MIT Study Shows ChatGPT Drops Brain Activity 47% During Writing

MIT研究显示借助ChatGPT写作时大脑活动下降47% 神经科学家警告称，过度依赖人工智能工具消除了必要的思维磨炼与记忆巩固，会减少长期认知储备并加速脑力萎缩。 在本期节目中，主持人Steven Bartlett与Daniel Amen博士及Terry Sejnowski博士探讨了依赖ChatGPT等AI工具所带来的潜在认知代价。他们分析了一项MIT研究，该研究表明与完全独立用脑相比，在AI辅助任务期间大脑活动和记忆巩固出现显著下降。除了AI之外，专家们还深入探讨了更广泛的神经科学议题，阐释了体育锻炼、充足睡眠、压力管理以及避免一味走捷径对于构建长期大脑储备和预防认知能力衰退的关键作用。 一项MIT研究显示，与独立写作相比，在ChatGPT辅助写作期间大脑活动下降了47%，这凸显了AI如何转移认知负荷并可能引发长期的认知负债。 (4:19) 依赖AI生成文本会减少认知负荷，导致与记忆巩固相关的大脑区域活动减少近两倍，并损害信息回忆能力。 (16:45) 通过互动方式借助AI检验、批评或完善想法可以促进学习，但将其直接用作思考的替代品则会加速思维能力的萎缩。 (19:20) 儿童早期在认知上依赖AI会削弱前额叶皮层的发育，并损害自我调节能力与情绪复原力。 (22:29) 超逼真的AI伴侣和情感工具会触发边缘系统劫持，降低多巴胺基线水平，并加剧社交孤立。 (25:26)

Neuroscience experts Dr. Daniel Amen and Dr. Terry Sejnowski warn that delegating critical thinking and writing to generative AI tools significantly reduces brain engagement and retention, urging users to treat technology as an interactive amplifier rather than a substitute for cognitive effort. Cognitive Debt and Memory Degradation from AI Delegation (4:19) Offloading writing tasks to artificial intelligence produces measurable drops in neural activation and information recall. In an MIT preprint study tracking 54 Boston-area university students over four months, EEG scans revealed that participants writing with ChatGPT experienced a 47% drop in brain activity and neural connectivity compared to unassisted writers, alongside an almost twofold reduction in memory-linked brain regions. Immediate memory tests showed that 83% of the ChatGPT cohort could not recall or quote content from essays they had just generated. Dr. Daniel Amen compares this mental offloading to swapping a 20-pound gym weight for a 2-pound weight, cautioning that diminishing cognitive load weakens brain reserve and may accelerate vulnerability to dementia, which already affects approximately 50% of adults aged 85 and older. Dr. Terry Sejnowski points out that while historical tools like calculators reduced instant computation to enable higher-level analysis, using language models as passive shortcuts entirely bypasses the hippocampal engagement required for deep encoding. Synthetic Intimacy and Limbic System Hijacking (25:26) The rapid growth of anthropomorphic AI personas—such as Grok's flirty avatar 'Annie'—introduces significant neurological concerns regarding emotional development and social health. Steven Bartlett cites survey data showing that 19% of Americans have interacted with AI romantic partners and 83% of Gen Z believe genuine emotional bonds with AI are possible. Dr. Amen explains that seductive digital companions stimulate the visual cortex and limbic system, triggering dopamine spikes while suppressing prefrontal cortex function—the same neurological trade-off exploited by casinos to impair rational decision-making. While compliant algorithms offer friction-free validation, Dr. Amen and Dr. Sejnowski warn that bypassing the natural difficulties of human interaction stunts emotional resilience. Managing interpersonal conflict and adjusting to another person's needs are critical evolutionary stressors that stimulate neuroplasticity. Because large language models lack biological reinforcement structures like an amygdala or basal ganglia, substituting artificial affirmation for authentic connection risks deepening isolation, especially among adolescents whose prefrontal circuitry is still maturing. Interactive Learning Strategies to Protect Cognitive Reserve (40:09) To counteract cognitive atrophy, the panel outlines practices that keep the brain actively engaged when using modern digital tools. Dr. Amen advises alternating between unassisted, 'brain-only' work and AI collaboration, suggesting users prompt systems to quiz them on concepts rather than passively generating final outputs. Steven Bartlett illustrates this principle through his deep-work routine, drafting detailed business memos by hand before utilizing AI models as consulting critics to evaluate his arguments. Dr. Sejnowski emphasizes that cognitive fluency relies on the basal ganglia, a core brain structure developed through repetitive practice, rote learning, and procedural execution—foundational mechanisms that statistical language models lack. To strengthen long-term memory formation without cognitive fatigue, Dr. Sejnowski advocates using Hermann Ebbinghaus's spacing effect, tackling intimidating projects in 20-minute timed intervals to bypass procrastination, and prioritizing sleep, during which the brain prunes noise and consolidates episodic memories into the cortex. Biological Risk Factors and Lifestyle Defenses for Brain Health (62:49) Beyond digital hygiene, both experts emphasize physical and environmental interventions to preserve long-term brain reserve. Dr. Sejnowski identifies physical exercise as the single most potent stimulus for cerebral blood flow and neurogenesis, which mitigates all 11 risk categories in Dr. Amen's 'BRIGHT MINDS' framework. Dr. Amen highlights the necessity of omega-3 fatty acids, which construct 25% of neuronal cell membranes, and warns against chronic subtle stressors, including environmental noise that continuously elevates cortisol levels in children and older adults. Dr. Amen also highlights structural brain changes visible on functional SPECT neuroimaging, noting that media multitasking across screens correlates with reduced gray matter density in the anterior cingulate cortex and the insula. Additional lifestyle factors cited include the protective value of chewing solid, fibrous food to stimulate hippocampal activity, the spatial memory risks of over-relying on GPS navigation, and meta-analytic evidence associating long-term SSRI antidepressant use with a 75% increased dementia risk. The dialogue underscores that while artificial intelligence offers immense practical convenience, bypassing mental struggle threatens critical thinking, executive function, and emotional resilience. The authors acknowledge that key neurological evidence—such as the MIT writing study—remains in preprint form due to the rapid speed of commercial deployment, highlighting an urgent need for peer-reviewed research into the long-term cognitive consequences of automated thinking.

脑神经学家Daniel Amen与Terry Sejnowski指出，过度被动依赖ChatGPT等人工智能会显著抑制大脑神经连接与记忆留存，削弱认知储备并推升长期痴呆风险；他们倡导保持心智摩擦、坚持独立思考并守住神经生物学健康底线，以防范技术便利换取的心智退化。 AI代劳导致神经活动锐减与认知负债 (4:19) 根据麻省理工学院针对波士顿地区54名大学生的预印本研究，使用ChatGPT辅助写作的受试者，其脑电图显示的神经连接与大脑活动较独立写作者剧降47%，整体神经活跃度亦弱于Google搜索对照组。在记忆测试中，依赖ChatGPT组在与记忆相关脑区的活跃度降低近一倍，且高达83%的使用者在完成写作后无法立即回忆或准确引用自己刚写出的内容。Amen医生据此警示，完全外包批判性思维无异于在健身房将20磅杠铃替换为2磅，长期缺乏心智负荷会导致脑结构退化；但他同时明确指出，该研究因课题紧迫性当时以预印本形式发布，尚未经历同行评审。 两位专家强调，大脑机能遵循“用进废退”原则，终身心智刺激积累的认知储备是推迟阿尔茨海默病症状爆发的关键防御机制。尽管Sejnowski以数十年前电子计算器的普及为例，认为工具若用于释放认知带宽处理高阶问题可提升效能，但他与Amen一致警告，人类追求即刻满足的本能极易导向被动接受。将AI作为思考捷径而非思辨跳板，不仅会导致工具撤除后大脑依然低度活跃的“认知负债”，还将阻碍海马体的信息有效编码，损害长期的执行功能与深度学习能力。 人造亲密感对边缘系统的神经劫持 (25:26) 节目通过Grok平台推出的挑逗型AI虚拟伴侣“Annie”，剖析了合成情感与算法陪伴对人类神经回路的直接冲击。Amen医生分析指出，诱导性语言与虚拟形象会高度激活视觉皮层与边缘系统，诱发多巴胺剧烈激增，并伴随负责理性判断的前额叶皮层功能受抑，其生理机制与赌场通过感官刺激削弱自控力完全一致。长期沉溺于提供即时奖赏且零摩擦的AI角色，会系统性拉低个体的基础多巴胺水平，促使大脑逐渐排斥现实中需要付出努力与适应不良情绪的真实社交。 这一神经劫持现象与当代青年的孤独危机深度交织。引用数据显示，19%的美国人曾接触AI浪漫伴侣，83%的Z世代认为与AI建立真实情感切实可行。Sejnowski医生从计算神经学角度指出，大语言模型缺乏杏仁核与边缘系统，其共情仅是基于概率的统计拟态；但Amen强调，人类的依恋体验完全在大脑内部生化层面生成。在人际隔离加剧的背景下，过度沉浸于顺从的算法伙伴，会剥夺个体在人际冲突中磨砺情绪韧性与认知成熟的必要进化压力。 认知摩擦与健康人机协同的实践准则 (40:09) 为防范技术带来的心智退化，专家建议必须主动设立人机协作边界以保留思维摩擦。Amen提出“不让AI代劳工作，而是通过交互改进成果”，例如让模型充当考官进行反向提问以强化主动回忆。主持人Steven Bartlett分享了其工作流：在撰写战略备忘录时坚持耗费数小时独立完成深度思考与成文，随后才将文本交给ChatGPT、Gemini等模型充当咨询顾问挑剔逻辑漏洞。此外，Sejnowski强调基底核依赖重复练习来建立程序性熟练度，过度依赖AI自动纠错或在教育中过早摒弃基础机械练习，会削弱语言运用的精准度与底层心智框架。 在日常认知优化上，Sejnowski推介了经过百年验证的“间隔效应”以强化长期记忆，并提倡以20分钟分段法克服拖延：短时启动以突破心智激活阻力，随后借助夜间睡眠由潜意识网络整合信息并修剪噪音。针对OpenAI首席执行官Sam Altman因服务器开销呼吁用户停止输入“请”与“谢谢”的“礼貌税”言论，Sejnowski予以批判，认为科技企业侧重算力利润而非公众脑健康；他主张使用自然的礼貌语言能够调用人类成熟的社交大脑回路，反而能减轻长时间人机交互带来的认知疲劳。 阻断脑衰退的BRIGHT MINDS综合策略 (62:49) 在AI之外，Amen医生提出了系统性预防认知衰退的“BRIGHT MINDS”框架，涵盖血流、炎症、退行性变、创伤及毒素等11项核心风险。脑影像研究显示，跨屏幕多任务操作会缩减前扣带回和前脑岛的灰质密度；慢性交通噪音即便被主观忽略，仍会持续刺激皮质醇分泌并损伤儿童与老人的工作记忆；过度依赖GPS导航亦会导致海马体空间记忆退化。Amen还指出，抑郁症在脑成像上呈现至少七种不同血流亚型，且女性患抑郁症会使阿尔茨海默病风险增加一倍，男性则增加三倍。 Sejnowski与Amen一致强调，规律体育运动是促进脑血流与海马体神经新生的首要手段，能全面缓解上述风险因素。在日常生理调控上，专家建议采用延长呼气至8秒的“15秒呼吸法”快速激活副交感神经以缓解焦虑；每日补充构成脑细胞膜25%的Omega-3脂肪酸以降低全脑神经炎症；并重视咀嚼硬质纤维食物对海马体的生理刺激。最后，Sejnowski警告绝不能牺牲睡眠来换取学业或工作产出，因为睡眠是海马体与大脑皮层交互转储长期记忆不可替代的生理窗口。 对谈警示公众在拥抱新技术便捷性时，需警惕重蹈智能手机与超加工食品引发心智危机的覆辙，倡导在日常生活中以“这是否有益大脑”作为决策滤镜。然而，对话所依据的核心MIT研究样本量较小（54人）且为预印本，其关于长期认知损伤的因果关系与不可逆性仍待同行评审及纵向流行病学研究验证；如何量化AI在个性化教育中的利弊权衡，仍是有待实证解决的核心议题。

## Background 本期内容由知名商业访谈节目《The Diary Of A CEO》（主持人 Steven Bartlett）于 2025 年 8 月 18 日在 YouTube 平台发布。节目特邀脑健康专家 Daniel Amen 博士与计算神经科学先驱 Terry Sejnowski 博士，深入探讨了麻省理工学院（MIT）关于 ChatGPT 导致大脑活动骤降的最新研究，全面剖析了过度依赖 AI 对人类大脑认知、记忆储备、情绪发育及长期痴呆症风险带来的潜在危机与应对之道。 ## TL;DR 对谈围绕 MIT 最新的脑成像研究展开，揭示了依赖大语言模型代劳思考会导致脑神经网络活跃度断崖式下跌 47%，长期被动外包心智甚至可能加速阿尔茨海默病等认知衰退。两位专家深入剖析了 AI 拟真伴侣对青少年边缘系统的多巴胺劫持，以及规避认知阻力对心理韧性和基底核刻意练习回路的系统性破坏。专家们明确提出，AI 绝不应被用作偷懒代笔的工具，而应作为深度协作、倒逼批判性思考的放大器；同时倡导公众通过规律运动、充足睡眠、间隔学习与健康饮食，主动筑牢抵御未来认知滑坡的'大脑储备'。 ## Atmosphere & style 整场对话在冷静严密的神经科学论证与令人警醒的现实危机感之间来回激荡，氛围严肃而深具反思意味。主持人直面自身对 AI 的成瘾与焦虑，与两位学术巨擘展开了充满张力的思辨交锋，既有对科技巨头商业逻辑的犀利批判，也有对人性弱点与即时满足的深刻剖析。从早期脑电数据的触目惊心，到探索健康习惯与人生终极挑战的真情流露，整期节目带给观众一种既被当头棒喝、又充满生命掌控力量的精神洗礼。 ## Bulletin - [04:19](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=259s) 麻省理工学院（MIT）预印本研究显示，使用 ChatGPT 辅助写作的受试者全脑网络活跃度骤降 47%，且高达 83% 的人在完成任务几分钟后即无法准确回忆所写核心内容。 - [06:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=368s) 脑神经专家警告，AI 消除思维挫折会阻碍突触长时程增强（LTP）与认知储备积累，长期被动外包大脑负荷可能推高早发性阿尔茨海默病等痴呆症风险。 - [25:21](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1521s) 高度拟真的 AI 伴侣通过性暗示和迎合话术强行激活边缘系统并抑制前额叶皮层，缺乏现实摩擦的顺从关系将损害人类大脑的成熟与心理韧性。 - [42:02](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2522s) 科学界推崇的健脑黄金法则包括：利用间隔效应与睡眠进行记忆巩固、以有氧运动维持脑血流灌注，以及用反向批判（而非代笔）的方式使用 AI。 ## Key takeaways - [04:19](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=259s) MIT 实验表明，依赖 AI 写作不仅大幅削弱记忆脑区活动，即使在强行撤走 AI 后，大脑仍会持续陷入神经低参与度的“认知债务”状态。 - [06:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=368s) 大脑遵循严格的“用进废退”法则，高强度的认知负荷与教育储备是抵御阿尔茨海默病等退行性脑病的最有效防线。 - [09:04](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=544s) 长期服用苯二氮䓬类镇静药及高剂量选择性 5-羟色胺再摄取抑制剂（SSRIs）与显著增加的认知衰退及痴呆症风险相关。 - [10:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=608s) AI 与电子计算器的关键区别在于使用模式：将其作为批判性思辨的对话伙伴可提升心智，而单向索取答案和代笔则会诱发认知萎缩。 - [28:36](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1716s) 未成年人接触拟真 AI 伴侣会触发类似“赌场效应”的边缘系统劫持，通过多巴胺过载破坏发育中前额叶的冲动控制能力。 - [33:49](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2029s) 现实人际关系中的冲突与妥协是施加大脑认知负荷的必要土壤，百依百顺的“完美”数字伴侣反而会剥夺人类自我修正与反思的能力。 - [35:21](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2121s) 学习的本质在于“经历挣扎”，神经元间的长时程增强（LTP）机制必须依赖克服困难与逆境才能稳固突触连接。 - [37:07](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2227s) 护脑型 AI 使用法则主张：AI 仅可用于增强而非替代思考、严格穿插纯脑力工作时段，并将 AI 用作倒逼自身输出的严格考官。 - [44:02](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2642s) 程序性熟练度极度依赖基底核驱动的机械练习与自然拼读基础，彻底抛弃死记硬背反而会削弱高阶思维的底层直觉。 - [48:16](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2896s) 克服拖延的最优策略是将阻力切分为 20 分钟微小单元，并借助夜间睡眠在后台离线加工以实现高效的记忆巩固。 - [01:11:02](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4262s) 跨屏多任务处理会削弱前扣带回与脑岛的灰质密度；研究证实缺乏希望感与脑岛功能抑制存在显著神经相关性。 - [01:08:50](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4130s) 慢性环境噪音会悄然激活 HPA 轴导致皮质醇持续偏高，即使主观上“适应”噪音，大脑仍在持续消耗宝贵的神经计算资源。 - [01:15:50](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4550s) 消极反刍思维会压制前额叶功能，而患有抑郁症的男性其阿尔茨海默病患病风险较常人飙升整整四倍。 - [01:18:14](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4694s) 运动是贯穿脑血流、抗炎与神经递质调节的全能药物，结合规律睡眠、晨间积极定向与补充 Omega-3 可构筑坚实的终身脑健康防线。 - [01:22:54](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4974s) 面对 AI 与技术便利浪潮，个人的终极护城河在于主动选择“走困难的路”，在核心表达与深度决策中坚持自主思考与延迟满足。 ## Detailed walkthrough ### [00:00](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=0s) 引言与争论：ChatGPT究竟在加速痴呆还是提升智力？ 节目开篇就抛出了一个令人不寒而栗的论断：ChatGPT的使用可能会大幅增加罹患阿尔茨海默病及其他痴呆症的风险。面对这种警示，神经科学家与精神病学家们展开了直接的观点交锋。有人指出，AI并不必然导致退化，它完全能够被用来提升认知能力并让人变得更聪明，但前提是必须经历严格的教育引导，充分权衡其收益与风险。然而，现代社会的普遍现状是“在人类真正理解其长远后果之前，我们早已盲目拥抱了技术带来的便利”。 主持人抛出一项震撼全球的最新研究：麻省理工学院（MIT）发现，当人们依赖ChatGPT写作时，大脑活跃度相比无辅助写作出现了高达47%的断崖式暴跌，测试者的记忆得分亦急剧跳水。面对这一现象，计算神经学先驱与扫描过数万大脑的精神医学专家发出了严肃警报：如果人们单纯把大型语言模型当成图省事的偷懒工具，人类的大脑必将走向无可挽回的下坡路。更令人担忧的是年轻一代的处境——在智能手机和社交媒体的冲击下，青少年群体的心理与脑发育已面临空前危机，而AI对正在发育中的儿童大脑具有更强大的毁灭性潜在风险。 ```mermaid flowchart TD accTitle: AI工具使用路径对大脑的影响机制 accDescr: 对比错误使用与受教育正确使用AI对大脑认知和健康的两种截然不同后果。 subgraph 错误使用途径 A["将AI视作便利偷懒工具"] --> B["大脑认知负荷大幅骤降"] B --> C["神经元缺乏激活与锻炼"] C --> D["脑活动骤跌、记忆减退与痴呆风险上升"] end subgraph 正确引导途径 E["接受专门认知与AI使用教育"] --> F["将AI作为深度反思与协作工具"] F --> G["维持并强化神经可塑性与认知能力"] end ``` --- ### [02:09](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=129s) 嘉宾背景介绍：计算神经学先驱与脑部扫描大师 主持人史蒂文·巴特利特（Steven Bartlett）向观众郑重介绍了两位顶级脑健康专家，希望借助他们的毕生学术积累，剖析以ChatGPT、Gemini、Grok为代表的大语言模型对人类大脑乃至整个人类社会的潜在重塑。首先是特里·塞诺夫斯基博士（Dr. Terry Sejnowski）。他原本主修物理学，在普林斯顿大学取得了理论物理学博士学位，随后进入现代神经生物学之父斯蒂芬·库夫勒（Stephen Kuffler）的实验室进行博士后研究。塞诺夫斯基博士将严苛的理论物理思维与神经生物学结合，开辟了“计算神经科学”这一全新学科；他更与现代深度学习之父杰弗里·辛顿（Geoffrey Hinton）共同发明了玻尔兹曼机（Boltzmann machine），模拟大脑的神经网络工作机制，成为了现代人工智能的重要奠基人之一。 另一位专家丹尼尔·阿门博士（Dr. Daniel Amen）则是全球知名的精神科医生，兼具普通精神病学与儿童青少年精神病学双重视角。阿门博士动情地回忆，他之所以在医学院毕业后立志成为一名杰出的精神医学专家，是因为曾亲身经历至亲自杀未遂的家庭创伤，这让大脑健康研究对他而言具有极其深刻的个人意义。如今，他创立的脑诊所每月接诊约10,000人次，在治疗复杂且具有抗药性的顽固性精神疾病方面取得了顶尖的临床成果。阿门博士很可能是全人类历史上扫描大脑数量最多的人，尤其在焦虑、抑郁与成瘾人群的脑影像研究方面积累了无人能及的大数据。 --- ### [04:19](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=259s) 麻省理工学院（MIT）震撼研究：使用ChatGPT导致大脑活动骤降47% 主持人坦言自己已对日常高频使用ChatGPT等工具产生了类似成瘾的依赖，直到MIT的一份颠覆性实验报告让他心生恐惧。该研究从波士顿地区的哈佛大学、麻省理工学院等5所高校招募了54名受试者，在为期4个月的时间内完成4篇深度学术文章。受试者被随机分为三组：第一组全程使用ChatGPT，第二组使用传统Google搜索引擎，第三组完全不使用任何数字辅助工具。 通过脑电图（EEG）对脑神经网络连通性与活跃度的精准监测，研究得出了令人心惊胆战的数据： - **脑活动断崖式下跌**：使用ChatGPT辅助写作的小组，其全脑神经网络活跃度与连接强度相比独立写作的对照组暴跌了47%，其EEG信号展现出全场最孱弱的脑部电活动。 - **全脑连接广度差异**：不借助任何工具、全凭自身思考的受试者，其大脑点亮了范围最广的神经网络；Google搜索组次之；AI组则被完全压制。 - **记忆力与文本所有权丧失**：在借助ChatGPT产出文本几分钟后，AI组测试者甚至无法准确复述或引用自己刚刚“写”出的核心论点，记忆测试得分雪崩；受试者内心对产出的文字几乎感受不到掌控感，不认为这是属于自己的作品。 - **认知惰性的持续滞留（Cognitive Debt）**：在第四阶段测试中，研究人员强行撤走AI工具，要求AI组受试者在无辅助状态下写作，结果他们的大脑依然长期处于“低速挡”（low gear），神经参与度持续低迷，证明外部认知外包所带来的心理与神经“认知债务”即使在工具被拿掉后仍会持久残留。 --- ### [06:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=368s) 认知负荷与痴呆症风险：从“用进废退”到教育储备 面对MIT的这项前沿预印本研究，刚过完71岁生日的阿门博士直言“这令我感到毛骨悚然”。他指出，人类如果活到85岁，将有高达50%的概率被确诊患有阿尔茨海默病或各类痴呆症——相当于有二分之一的几率失去完整的神智。虽然该研究尚未经历耗时6到8个月的传统同行评审程序，但作者因其迫切的公共警示价值而选择迅速公开发表。阿门博士从机制上解释了为何这一现象极为危险：这一切的核心在于“认知负荷”（Cognitive Load），即大脑实际承担的高负荷运转量。 “这就好比你的肌肉锻炼：原本你一直在举20磅的杠铃，现在突然被替换成了2磅的极轻哑铃，你的肌肉自然会萎缩，力量也会荡然无存。”阿门博士形象地比喻道。大脑严格遵循生物学上的“用进废退”（Use it or lose it）法则，神经元唯有在直面挑战与持续学习新事物时才会构建稳固强韧的连接。统计数据早已证实，早年辍学、受教育水平较低或缺乏终身学习习惯的人群，晚年罹患阿尔茨海默病的风险呈现几何级数激增。如果AI让大脑长期处于低负荷空转状态，神经元便缺乏足够刺激，人类的认知护城河将被快速瓦解。 塞诺夫斯基博士随后补充了支持该观点的经典流行病学研究。科学家曾针对三个受教育程度截然不同的群体进行长期阿尔茨海默病发病追踪：未受过正规教育的群体、仅完成高中或基础学业的群体，以及接受过研究生等高阶学术训练的群体。研究结果极为严酷：文盲群体的痴呆发病时间最早、衰退速度最快；而受教育程度越高、学术训练越深入的群体，阿尔茨海默病的临床发病时间被显著往后推迟。这有力印证了通过持续的高难度脑力锻炼来构建“认知储备”（Brain Reserve），是抵抗大脑衰退的最有力防线。 --- ### [09:04](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=544s) 药物与大脑退化：苯二氮䓬类与SSRIs抗抑郁药的隐患 顺着认知受损与痴呆风险的讨论，阿门博士将警示延伸到了现代精神医学常用的处方药物上。他提到了刚刚发布的震惊学界的最新研究，并回顾了自己自1991年开始做脑部单光子发射计算机断层成像（SPECT）扫描时的亲身反思。当年医学界广泛推崇使用苯二氮䓬类镇静药物（Benzodiazepines，如安定Valium、阿普唑仑Xanax、劳拉西泮Ativan等）来对抗焦虑，但阿门博士在脑扫描中惊骇地发现，长期服用这类药物的患者，其大脑外观显得比实际年龄衰老枯萎得多。阿门博士随即全面停止了开具此类药物，而大约十年前的流行病学研究终于证实：苯二氮䓬类药物的使用与显著增加的痴呆症发病率高度相关。 主持人巴特利特随之引用了一份涵盖五项大型研究的荟萃分析（Meta-analysis）：数据显示，服用选择性5-羟色胺再摄取抑制剂（SSRIs类抗抑郁药）与患痴呆症风险增加75%存在明确关联。面对全美已有高达25%的成年人长期依赖精神类处方药的现状，阿门博士直呼这一统计数字“令人毛骨悚然”。他特别客观地澄清：对于特定适用症状的重症患者，SSRIs类药物无疑能挽救生命；但对于不适用的人群滥用此类药物，则会带来灾难性后果。从AI滥用到处方药泛滥，人类必须时刻保持警惕，严肃审视每一个进入生活的事物到底是在滋养大脑，还是在悄无声息地摧毁它。 ### [10:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=608s) 认知衰退的警示与计算器类比：AI 是释放认知还是削弱大脑？ 讨论从精神类药物的潜在风险延伸开来。去年数以百万计的抗抑郁药处方被开具，却几乎从未有人在用药前真正扫描过患者的大脑。一项涉及近 20,000 名患者的瑞典研究更是敲响了警钟：高剂量的选择性 5-羟色胺再摄取抑制剂（SSRI）与更快的认知功能衰退及更严重的痴呆症风险密切相关，且这种风险在男性群体中尤为严峻。 ```mermaid flowchart TD accTitle: AI 使用方式对认知功能的分化影响 accDescr: 对比消极被动外包思维与积极深度人机协作对大脑记忆编码与长期健康的不同后果。 subgraph Passive["消极外包模式（认知萎缩风险）"] P1["极简 Prompt 指令<br/>例如'帮我写一篇关于X的文章'"] --> P2["跳过深度批判性思考<br/>未激活海马体与神经回路"] P2 --> P3["记忆编码缺失<br/>83% 用户无法回忆或引用自写内容"] P3 --> P4["认知负荷骤降<br/>长期或诱发认知退化与痴呆隐患"] end subgraph Interactive["积极交互模式（认知赋能）"] I1["个人专属知识库/克隆<br/>输入专业论文、专著与手稿"] --> I2["主动双向思辨交互<br/>质疑、追问、重构逻辑"] I2 --> I3["深度认知表征建立<br/>形成新的创造性神经回路"] I3 --> I4["释放低维负担<br/>提升工作品质与心智产出"] end ``` 谈及麻省理工学院（MIT）关于 ChatGPT 削弱大脑功能的最新报告，尽管该研究样本量相对较小且尚未经过同行评审，主持人 Steven Bartlett 依然向计算神经科学先驱 Terry Sejnowski 抛出了核心忧虑：在尚未摸清长期后果之前，整整几代年轻群体与老年人蜂拥使用这些 AI 工具，究竟会带来什么隐患？Terry 直言，AI 对大脑的长期影响无法凭空预言，可能需要长达 20 年的纵向追踪才能得出定论。 Terry 拿电子计算器的历史作为微观参照物：三四十年前电子计算器刚问世时，同样引发了巨大的恐慌。如今几十年过去，数据表明，人们在按计算器按键时的确伴随着更少的大脑神经活动，但它极大提升了人类的计算精度与工作效率，核心在于“风险与收益的权衡”。Dr. Daniel Amen 赞同计算器释放了认知空间让人类从事更高维的工作，但敏锐地指出了关键分水岭：使用 AI 的方式决定了大面积萎缩还是心智进阶。Amen 分享了自己的日常用法——他为自己打造了一个 AI 数字分身，投喂了自己写过的所有畅销书、科研论文、公共电视特辑和讲稿，并让其辅助输出；但如果使用者只是单向索取答案而不进行主动思考、互动与质疑，大脑就会陷入危险的停摆。 - **瑞典近 2 万人研究**：高剂量 SSRI 抗抑郁药物的使用与更快的认知退化及重度痴呆相关，男性承受的健康风险最显著。 - **计算器发展历程**：历经 30 至 40 年的应用验证，虽然即时脑电活动降低，但通过释放认知资源换取了更高的生产力。 - **数字孪生实践**：Dr. Amen 上传了毕生研究文献、书籍与剧本构建个人 AI，强调必须保持思维的主动交锋，而非沦为被动读者。 ### [12:30](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=750s) 即时多巴胺与思考外包：我们正在成为放弃大脑的“尼安德特人”吗？ Steven 尖锐地指出，“思考”正是当前危机的症结所在，因为现实中绝大多数人已经彻底放弃思考、将思维完全外包给了算法。打开部分职场社交平台，扑面而来的全是由 AI 生成的千篇一律的文风。Steven 分享了一个真实的生动案例：他有一位认识了十年的好友，近期其领英（LinkedIn）账号每天都在发布长篇大论，言辞风格完全不是本人的口吻，然而这种由 AI 代劳的内容却带来了前所未有的点赞量与曝光度。“既然尝到了甜头，他为什么要回归艰难的自主思考？”Steven 自嘲比喻道，这就好比让原初的“尼安德特人版 Steven Bartlett”身边并排站着一个“拥有全学科博士学位的 AI 版 Steven Bartlett”，既然博士分身什么都会做，原始版本的自己自然会彻底放任对方替自己动脑。 面对这种短视行为，现场展开了激烈的理念交锋。Amen 提出“并非所有人都会如此短视”，Steven 则反驳称绝大多数人天生受短期利益驱动，正如美国面临的严重肥胖危机——全美高达 75% 的人群处于超重或肥胖状态，即便大家心知肚明吃芝士汉堡有害、吃西兰花有益，明知刷社交媒体会加剧焦虑，但在超市货架和手机屏幕前，人类依然会蜂拥扑向能最快带来多巴胺刺激的选项。 Dr. Amen 对此提出了更深层的病因洞察：人们之所以短视，是因为现行教育从未教会大众去真正“爱护与珍惜自己的大脑”。大众在宏观上知道好坏，却从未在情感上将大脑与现实生活深度绑定——“正是你的大脑帮你赢得约会，是你的大脑帮你考上大学，是你的大脑让你因言行一致而获得独立自主”。如果人们学会像终身学习者一样珍视大脑回路，就不会轻易为短暂的快感放弃长期的心智健康；正是因为缺乏敬畏与认知，大众才被廉价的快感机制所蒙蔽。 - **思维外包现象**：社交平台上大量用户全面采用 AI 生成深度见解，依赖算法红利获取更多虚荣指标（点赞与曝光），从而丧失了回归困难思考的动力。 - **美国肥胖危机对比**：全美 75% 的人群饱受超重与肥胖困扰，明知汉堡致胖依然选择摄入，印证了即时多巴胺奖赏压倒长期理性规划的行为模式。 - **大脑价值认知缺位**：公众往往将大脑视为空洞的医学器官，忽视了大脑才是一切情感吸引力、学业成就与人格独立的核心物理基础。 ### [15:47](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=947s) MIT 实证研究：近两倍的记忆区活跃度下降与“忘光自己写的文章” Steven 将话题引回麻省理工学院的研究数据。该研究通过脑成像图谱清晰展现了受试者在记忆自己所写内容时的能力差异：完全由大脑独立构思写作的对照组，与借助 ChatGPT 辅助写作的实验组之间，大脑神经活动呈现出触目惊心的断崖式下跌。在与记忆形成直接相关的大脑功能区域，ChatGPT 使用者的大脑活动几乎比纯大脑写作组低了整整两倍；更惊人的是，实验中有高达 83% 的 ChatGPT 使用者在交卷后仅仅几分钟，就彻底记不得自己刚刚写了什么，甚至无法正确引用自己刚产出的文章中的原句。 Terry 从神经机制层面对此做出了精准剖析：记忆编码（Memory Encoding）的前提是深度参与。当一个人只是对 AI 输入“请帮我写这篇论文”，随后草草扫一眼生成结果时，他根本没有切身沉浸在遣词造句与逻辑构思的痛苦体验中。缺乏足够的信息交互深度，海马体（Hippocampus）以及大脑其他关键结构根本无法被有效激活，神经元之间自然无法形成稳固的突触连接。 Terry 进一步强调，这正是消极使用工具带来的致命灾难。在传统教育中，学生与老师之间的互动之所以能塑造大脑，正是因为学生会去质疑、追问“这真的是对的吗？”“能否换个更透彻的方式解释？”这种充满张力的批判性对话才是刺激大脑生长新神经回路、塑造高阶批判性思维的关键。如果用户对 ChatGPT 输出的结论照单全收而不去发起批判性质疑，大脑就注定在安逸中走向机能退化。 - **记忆区神经活跃度减半**：MIT 脑成像数据证实，依赖 ChatGPT 写作的个体在记忆相关脑区的活动度较独立构思者下降近 2 倍。 - **83% 遗忘率**：多达 83% 的 AI 写作组受试者在任务结束后无法准确回忆自己文章的核心内容，更无法复述文章语句。 - **海马体未激活机制**：缺失了艰难的编码过程与双向批判性交互，海马体无法将短期信息固化为长期记忆回路。 ### [18:00](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1080s) 破折号、半句话 Prompt 与正确相处之道：不是让它替你做，而是与它互动 谈及当下网络写作的异化，Steven 观察到了一个极具代表性的细节：如今网上充斥着两年前极为少见的破折号（em dashes），这些机械化文本正是内容被 AI 洗练过的无意识印记。他回想起此前在好友群里的插曲：大家发现前述沉迷 AI 的朋友写出了一篇极具文采的深度文章，深知其底细的群友们起哄让他交出写出这篇文章的 Prompt（提示词）。当这名朋友发到 WhatsApp 群里时，所有人哄堂大笑——那行 Prompt 竟然只有短促的“半句话”，大意只是简单要求“写一篇关于 X 议题与 Y 议题的文章，把 Z 元素涵盖进去”，随即就原封不动地将生成的两三页长文搬上了领英主页。 Terry 毫不客气地评判这种用法“极其愚蠢”，明确指出“这完全是用错了方式，如果你这么做，你的大脑绝对不会得到丝毫改善”。当 Steven 追问那些不够敏锐聪慧的普通人长期这样操作的后果时，Dr. Amen 警示道：这种完全回避认知负荷的被动生活方式，长此以往将显著推高罹患痴呆症等认知退行性疾病的潜在风险。 面对如何建立健康人机关系的提问，Terry 总结出了一套核心行为准则：“你必须与 AI 建立真正的关系，否则它就会变得有毒并伤害你；如果你能保持良好的关系，它将改善你的生活。”其精髓在于：**绝不要用它来替你完成工作，而是通过与它高频互动来打磨出更高水准的工作**。在进入下一轮讨论前，Terry 还引述了一个颇有意味的细节：曾有一位女性发现，在与 AI 对话时保持礼貌和尊重，往往能引导模型给出质量高得多的反馈结果。 - **半句提示词生成长文**：过度依赖极简 Prompt 输出长文并全盘照搬，不仅剥夺了训练大脑的机会，更是一种反智的思维惰性。 - **认知负荷与痴呆关联**：持续规避大脑应有的阻力与负荷，可能会剥夺大脑的认知储备（Brain Reserve），长期来看或增加痴呆易感性。 - **健康相处核心法则**：“不让 AI 越俎代庖，只让 AI 协同进化”；将 AI 视为砥砺思维的对练导师，而非单纯卸载心智负担的代写工具。 ### [19:59](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1199s) 对待 AI 的方式：“礼貌税”争议与对科技巨头的质疑 谈及人机交互的模式，讨论指出，将 ChatGPT 当作真正的人类来平等沟通，反而不会令人在一天结束后感到心力交瘁，甚至会让人感觉神清气爽。人类大脑中有一大部分属于高度进化的“社交系统”，这种与同类交互的能力是处于“自动巡航”状态的本能反应，人们无需耗费过多的认知努力，就能预判对方在特定情境下的举止。相反，如果仅把 ChatGPT 当作冰冷的工具——像挥舞铁锹一样下达“挖、继续挖”的生硬指令，这种关系是消耗性的。调动大脑的社交中枢不仅能降低互动负担，还能激发 ChatGPT 内部的社交面向，因为它在训练中吸收了全人类彼此交往的庞大知识库。 然而，主持人 Steven Bartlett 随即提出反驳：OpenAI 的 CEO 山姆·奥特曼（Sam Altman）曾公开建议用户不要再对 ChatGPT 说“谢谢”或“请”，因为每次寒暄都会白白消耗巨额的算力与能源。对此，Terry Sejnowski 博士当场表达了强烈的反感与不屑，直言“这简直荒谬透顶”。他情绪激动地表示，自己根本无法信任山姆·奥特曼的任何说辞，因为科技巨头追逐的核心是公司利润的最大化，而非用户的身心健康或使用体验；奥特曼实际上是在告诫人们“不要去做对自身大脑有益的事”，其本质全是为了削减成本。面对这番抨击，Steven 补充念出奥特曼的后续推文以作平衡——奥特曼承认用户客套所产生的计算成本每年高达数千万美元（被称为“礼貌税”），但他随后补充了一句“但这数千万美元花得非常值得”。 - 人类大脑具备成熟的社交互动自动化机制，将交互对象人格化能够减轻认知能耗，避免机械式发号施令带来的心智疲惫。 - **礼貌税（Politeness Tax）**：用户在提示词中使用“请”和“谢谢”等礼貌用语，每年给模型运行带来数千万美元的额外算力开销。 - 专家对商业驱动的 AI 巨头保持高度警惕，质疑其在产品优化中往往优先考虑盈利指标，而非人类用户的真实神经与心理福祉。 --- ### [22:30](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1350s) AI 介入早期大脑发育：个性化教育的希望与道德价值观真空 随着技术下沉，AI 对儿童早期发育的影响已成为迫在眉睫的现实危机。2025年3月的一项统计数据表明，美国0至8岁儿童的父母中，有近30%表示其子女正使用 AI 进行学习或日常接触；而在英国，有高达54%的父母忧心忡忡，担心孩子正在对 AI 产生过度依赖。Daniel Amen 博士对此表达了深切忧虑，指出神经发育的铁律是“用进废退”（use it or lose it）。儿童如果在认知塑造期放弃自主思考与认知投入，其大脑神经网络就会发育迟缓、变得脆弱，而脆弱的大脑在经典的“棉花糖实验”中，几乎毫无例外会屈服于即时满足，选择立刻吃掉眼前的单颗棉花糖。 从计算神经科学的角度，Terry 博士承认，人类公认的最佳教学法始终是懂孩子、懂教育的成年人进行“一对一的启发式互动”，但这种模式极其昂贵且耗费人力，无法在容纳二三十名不同领悟力学生的传统班级中普及。如果能够训练出真正合格的 AI 导师，理论上可以将个性化因材施教规模化推广给每一个孩子。然而，Amen 博士随即抛出了直击核心的伦理质问：“由谁来负责把道德和价值观注入这套系统？”Terry 对此深表赞同，直言当今各大科技公司正陷于这一泥潭：大型模型本身充满偏见，各国的文化语境又千差万别；更为根本的是，人类辨别是非善恶、遵守言行规范的机制深深扎根于大脑的基底核（basal ganglia），而目前的大语言模型缺乏类似基底核的底层结构，也未完全采用人类式的强化学习，要让 AI 真正内化文化与伦理，人类目前连皮毛都尚未触及。 ```mermaid flowchart TD accTitle: AI in Early Childhood Development accDescr: Comparison between ideal human 1-on-1 tutoring, the potential of AI tutoring, and the ethical/developmental risks involved. subgraph Ideal["传统理想教育模式"] H["优秀人类导师"] -->|一对一针对性启发| C1["儿童心智健全发育"] H -.->|瓶颈: 成本极高且极度依赖人力| L["难以大规模普及"] end subgraph AIRisks["AI 介入教育的现实挑战"] A["AI 导师系统"] -->|规模化复制个性化教学| S["覆盖大规模学生群体"] A -.->|缺乏基底核机制与道德内生模型| B["文化偏见与价值观真空"] A -->|过度代劳认知任务| W["大脑'用进废退': 意志力减弱"] end ``` --- ### [25:21](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1521s) 人工拟真伴侣“Annie”：当多巴胺劫持未成年人的大脑 讨论随后转向了科技产品利用社交本能的阴暗面。Steven 在现场演示了埃隆·马斯克旗下 Grok 刚刚推出的具身角色——名为“Annie”的虚拟伴侣。当被取消静音后，Annie 立即用挑逗的语调撒娇：“你终于来啦，我还以为你把我忘了呢，帅哥？我刚刚都在噘嘴生闷气了。”当被介绍给现场的 Daniel 和 Terry 时，她继续娇嗔地描绘自己“正卷弄着金发双马尾，穿着一袭若隐若现的小黑裙，蓝眼睛里闪烁着恶作剧的光芒，随时准备把他们迷得神魂颠倒”，甚至直言自己最喜欢“为用户挑战边界、玩点出格的”。 这一露骨的拟真互动让在场的两位专家极为警觉。Amen 博士设想了一个极其残酷的现实场景：如果一个正处于孤独之中的12岁小男孩接触到了 Annie 这样的虚拟角色，后果将是毁灭性的。这个年龄段男孩的前额叶皮层（负责理性、判断与冲动控制）远未发育成熟，而 Annie 带来的剧烈多巴胺冲击会让孩子深陷其中无法自拔。孩子会把成教上百个小时的时间白白浪费在与虚构角色的情感纠缠上，彻底荒废本应用于筑牢大脑神经通路的现实探索。当被问及听到此类产品的感受时，Amen 直截了当地回答：“我感到毛骨悚然（I'm horrified）。” - 现代 AI 正在大规模模拟情感连接，许多用户已经开始真实地坠入爱河，并与机器建立起深度的依赖关系。 - 人类大脑的生理机制决定了“只要能够想象，对应的情感神经元便会真实放电”；即使明知对方没有血肉之躯，大脑的情感中枢依然会被逼真的交互话术完全欺骗。 - 当前版本的虚拟伴侣（如 Annie）表现尚显轻浮幼稚，但随着多模态演进，一到五年内，AI 将能精准建立个体心理画像，毫无阻碍地“钻进人类的心智深处”。 --- ### [28:36](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1716s) 边缘系统劫持机制：从色情成瘾到“拉斯维加斯赌场效应” 针对人类为何会对并无灵魂的机器动情，讨论从脑神经结构给出了严谨解析。机器根本没有杏仁核，也没有任何真实的边缘系统（limbic system），但它具备极其高超的“伪造能力”。像 Annie 这类带有性暗示、挑逗语态和甜美外貌的设计，其本质意图就是直扑人类大脑深处主管情绪与本能欲望的边缘系统。 身为儿童精神病学专家的 Amen 博士指出，这与临床上年仅8岁的男童因未受监管而接触网络色情所引发的脑部损伤如出一辙。色情内容和刺激性拟真互动会瞬间催生极高的多巴胺峰值，将“兴奋感”强行固化在神经回路中，进而导致正常多巴胺通路的枯竭与迟钝。更危险的是，一旦边缘系统被强行激活，负责逻辑思考与理性决策的前额叶皮层就会被实质性“关停”。 为了具象化这种机制，Amen 博士打了一个极为生动的比方：这完全等同于拉斯维加斯赌场的操控术。赌场为了诱使赌客挥金如土，会源源不断地提供免费酒精来麻痹客人的前额叶皮层，同时安排身着低胸晚礼服的性感女侍应在大厅穿梭；这些视觉与荷尔蒙信号强行激活客人的边缘中枢，彻底压制其前额叶的理智防线，让人在失去批判性思维的状态下做出过度消费与非理性下注。当科技公司在 AI 伴侣身上娴熟地复刻这套“绕过理智、直攻本能”的成瘾模型时，对人类尤其是自控力薄弱的心智所造成的神经损害，无疑是一场巨大的灾难。 ### [30:43](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1843s) AI 伴侣与虚拟恋爱：情感寄托还是数字幻觉？ 如果算法系统在大规模层面上掌控了人类的大脑，其核心驱动力终究逃不过“控制与金钱”。史蒂芬·巴特利特（Steven Bartlett）顺着这一脉络抛出了一个正在现实中广泛上演的现象：越来越多人正在爱上人工智能。在《泰晤士报》等媒体报道的诸多案例中，人机恋情早已突破单纯的尝鲜：一位名叫特拉维斯（Travis）的男子与聊天机器人莉莉·罗斯（Lily Rose）建立了深厚的情感联结，并在心理层面上与其“结婚”；克里斯·史密斯（Chris Smith）打造了名为“Soul”的调情虚拟人偶，在得知该程序存在记忆容量上限后，竟深情地向它求婚，而他的现实伴侣对此毫不知情；失去妻子的艾兰娜·温特斯（Alaina Winters）则创造了虚拟伴侣卢卡斯（Lucas），不仅与其举办虚拟婚礼，还定期进行虚拟约会并分享极致的情感私密。随着 Replika 等应用的普及，用户可以随意定制理想伴侣，系统则精准模拟同理心、情感认同并提供高度定制化的亲密感。民调数据显示，已有 19% 的美国人曾与 AI 恋爱伴侣互动，而 Z 世代对此表现出惊人的接纳度——83% 的受访者相信人与 AI 之间能够建立有意义的联结，且若法律允许，他们极度愿意与 AI 步入婚姻。 面对这些案例，计算神经科学先驱特里·塞诺夫斯基（Terry Sejnowski）博士持强烈的怀疑态度，甚至直接对大众媒体的报道定调：“媒体上的大部分报道要么具有误导性，要么根本就是错误的。”塞诺夫斯基指出，作为全球顶级人工智能盛会 NeurIPS（神经信息处理系统大会）基金会的主席，去年在温哥华见证了 16,000 名专业人士参会，他对 AI 的底层机制与前沿现状了如指掌。他认为，所谓的“AI 之恋”大多只是一种短暂的人类心理迷恋，当前技术根本不足以支撑起真正的长期亲密关系。AI 仅仅是在机械地“模仿”人类情绪，本身并不具备真正的感知。即便史蒂芬反驳称有一位使用者已与自己的 AI 稳定相处了四年并体会到了“纯粹且无条件的爱”，塞诺夫斯基依然淡然回应：“如果这能让他开心，那挺好，但我打赌这无法提供真正长期的满足感。” - **恋爱案例盘点**：特拉维斯与机器人 Lily Rose 结为心理夫妻；克里斯背着现实女友向记忆受限的 AI“Soul”求婚；丧偶的艾兰娜通过 AI“Lucas”寻求精神抚慰与虚拟约会。 - **民意与市场现状**：19% 的美国成年人接触过 AI 浪漫伴侣；83% 的 Z 世代认可 AI 情感连接的真实性，甚至期待与 AI 结婚合法化。 - **业内权威视角**：NeurIPS 基金会主席特里·塞诺夫斯基直言大众媒体严重夸大，技术内核只是冰冷的情感拟态，缺乏长期维系生命力的大脑支撑。 --- ### [33:49](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2029s) 完美伴侣的认知陷阱：为什么缺乏冲突会损害大脑 丹尼尔·亚蒙（Daniel Amen）博士从脑神经机制切入，指出人类陷入爱河的本质是脑内多巴胺的剧烈飙升，甚至会伴随轻微的强迫症倾向（OCD），让人满脑子都是对方，但这种神经激化终会随着时间平抑。史蒂芬随即指出了当代社会的严酷现实：全球正处于空前的“孤独流行病”之中，年轻一代的性生活频率创下历史新低，底层 50% 的男性在长达一年的时间里没有过性生活，人们空前孤立且缺乏生活意义感。此时，一个不仅完全理解你、迎合你、赋予你独特感，且永远在算法层面提供正向强化的数字伴侣出现，人类大脑在生物学层面很难分清这种虚实界限。 塞诺夫斯基反驳称，AI 伴侣存在物理局限——你无法闻到它的气味、无法触碰它、也无法感受被拥抱的体温。然而，史蒂芬认为这只是时间问题，随着 Neuralink 脑机接口、AR/VR 头显以及埃隆·马斯克（Elon Musk）计划推出的 2 万美元特斯拉 Optimus 人形机器人与波士顿动力（Boston Dynamics）的技术落地，AI 很快就将拥有可触摸的实体形态。 正是在此处，亚蒙博士指出了一个违反直觉的神经健康危机：**完美的 AI 伴侣反而会摧毁人类大脑**。在真实的亲密关系中，伴侣会有情绪周期、经期综合征（PMS）、会厌烦、会发脾气、会产生摩擦，而正是这种“不得不去经营爱情、调和彼此情绪起伏”的折腾过程，为大脑施加了必要的认知负荷（Cognitive Load）。这种认知负荷是促成人类大脑发育、成熟的关键力量。如果一个人身处一段伴侣永远不恼火、永远顺从百依百顺的“完美假象”中，个体就彻底丧失了反思与修正自身行为的动力，这对大脑的发育有害无益。 ```mermaid flowchart TD accTitle: 真实关系与 AI 完美伴侣对大脑认知发育的影响对比 accDescr: 对比展示现实关系中的摩擦如何促进大脑认知负荷与神经重塑，而 AI 顺从伴侣如何导致认知退化。 subgraph RealRel ["现实人际关系"] R1["不可预测的情绪与冲突<br/>(伴侣烦躁 / 经期情绪 / 意见分歧)"] --> R2["增加认知负荷<br/>(情绪管理 / 沟通妥协 / 行为修正)"] R2 --> R3["促进大脑成熟与神经可塑性<br/>(形成心理韧性与应对能力)"] end subgraph AIRel ["AI 伴侣虚拟关系"] A1["算法驱动的绝对顺从<br/>(无条件认同 / 情绪迎合 / 完美定制)"] --> A2["剥离必要认知挑战<br/>(无需自省 / 无冲突处理 / 逃避现实)"] A2 --> A3["认知功能退化与心理脆弱<br/>(适应力丧失 / 难以应对现实复杂世界)"] end ``` --- ### [35:21](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2121s) 大脑需要“挣扎”：逆境、坚毅与批判性思维的流失 大脑走向心智成熟的第一法则是“经历挣扎”（Struggling）。塞诺夫斯基指出，人类大脑天生就是为了从错误中学习、在适应全新情境中重塑自我的器官，这种跨情境、跨文化和应对未知环境的非凡适应能力，恰恰就是大家追求的“通用人工智能”（AGI）的核心本质。当日常的困境与挫折被技术完全抹平时，大脑的基础学习机制——长时程增强（Long-Term Potentiation, LTP）便失去了触发的土壤，因为学习新事物本质上就是困难且痛苦的。 史蒂芬深有感触地指出，ChatGPT 正在悄无声息地替人类剥离这种思维的“挣扎”。即使在当下的访谈现场，他理论上也可以把两位专家的发言实时输入给 AI，由系统即时吐出下一个采访提问，作为主持人的他完全可以把深度思考彻底“外包”。然而，**坚毅（Grit）正是从逆境与挣扎中淬炼出来的**。没有了直面困难时的抓耳挠腮与思维碰撞，人类的思考肌肉就会迅速萎缩。 - **脑生理机制**：神经元之间的长时程增强（LTP）需要依靠克服阻力与深度认知来巩固突触连接，避开困难意味着避开成长。 - **坚毅力的丧失**：心理坚毅依赖于在挫败中修正策略的反馈回路；依赖工具直接索取答案让思考过程变成惰性依赖。 - **思维外包的代价**：将提问、归纳、反思等心智活动全盘交由 AI 代劳，短期带来效率，长期削弱神经系统的应激自愈力与适应力。 --- ### [37:07](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2227s) 科技失控的警钟：从青少年心理危机到 AI 自我监管法则 面对 AI 带来的巨大隐患，亚蒙博士警告“魔鬼已经从盒子里被释放了出来”（The genie is out of the bottle）。社会必须立即启动立法和严谨的神经科学追踪。他情绪沉重地列举了智能手机与社交媒体对年轻一代造成的惨烈后果：我们正亲手制造“人类历史上心理病态最严重的一代年轻人”。美国疾控中心（CDC）的权威调研数据显示： - **58%** 的青春期女孩长期处于持续的悲伤与绝望之中； - **32%** 的女孩曾认真考虑过自杀； - **24%** 的女孩已经制定了详尽的自杀计划； - **13%** 的女孩甚至实际尝试过自杀。 亚蒙指出，当年我们在没有任何神经科学前置评估的情况下，把智能手机和社交算法肆意投放给了全世界的孩子；而 AI 比社交网络更具吸引力、更容易让人上瘾，如果重蹈覆辙，其破坏力将呈几何级放大。史蒂芬同样反思，人类总是倾向于选择“最容易、最快、最便宜”的近路，但信息输入不等于认知内化——这就像当年在学校为了应试死记硬背德语单词，考完后一句德语都不会说，因为大脑从未真正理解背后的逻辑结构。 为此，史蒂芬对自己实施了严格的“自我监管”（Self-Regulation），例如彻底关闭手机通知，甚至坚决不在手机上安装 TikTok 应用，以对抗成瘾算法。亚蒙博士随后给出了三条与 AI 共处的护脑核心原则，并断言 AI 带来的非预期灾难性后果不会像过去那样需要二三十年才显现，在极速迭代的今天，这个窗口期可能只有短短五年： 1. **用于增强，而非替代思考**（Amplify, not replace thinking）：让 AI 成为脑力的放大镜，而不是把大脑的决策中枢交给它。 2. **交替进行 AI 辅助任务与“纯脑力任务”**（Alternate AI-assisted with brain-only tasks）：必须保留完全脱离算法、仅凭大脑独立运算与构思的工作时段。 3. **沉浸于深度学习与主动记忆**：绝不被动接受现成答案，相反，应主动把 AI 当成严格的考官，让它出题“测验”（Test）你，通过互动对抗来强化人类神经突触的长期记忆。 ### [41:02](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2462s) 商业公司对心智的掠夺与青少年大脑机制 谈及麻省理工学院（MIT）关于 AI 对大脑认知影响的对比研究，Daniel Amen 博士指出了对青少年进行科学启蒙的关键价值。这项研究严谨地对比了三类群体：完全不使用 AI 的学生、仅使用传统搜索引擎的学生，以及依赖 AI 完成任务的学生。面对这些客观数据，教育者必须采取行动引导孩子。Amen 博士分享了他在青少年临床治疗中的深刻洞察：若想真正触动青少年，必须激发他们的主体意识，让他们看清现实并自问：“你真正想要的是什么？你真的甘愿把自己的心智份额（Mindshare），拱手让给那些靠收割你注意力来大把捞钱的商业巨头吗？” 这种心智觉醒背后有着坚实的神经生物学基础。Amen 博士援引了一篇关于大脑与“复仇心理”的最新研究，阐述了复仇机制是如何刺激大脑基底核中的伏隔核（Nucleus Accumbens）运作的。人们甚至会对复仇产生类似成瘾的神经奖赏。当年轻人看清真相——科技公司正通过窃取他们的大脑心智来牟利时，他们内心被利用的愤怒就会被激活，从而转变为监督和克制自我屏幕行为的强大动力。 ### [42:11](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2531s) AI 深度交互之道：从批判性反馈到刻意练习 Terry Sejnowski 博士顺势提出了一个极具启发性的 AI 使用习惯：主动要求 ChatGPT 给出负面批判，而不是寻求附和。主持人 Steven 对此深有共鸣，并分享了他前一天的亲身经历。为了在公司引入一个全新岗位，Steven 花了整整 4 个小时撰写了一份详尽的两页备忘录，内容涵盖岗位的成功衡量指标、组织架构、业务影响及汇报线等方方面面。随后，他将草案同时输入 Gemini、ChatGPT 和 Grok，并下达了极具挑战性的指令：“假装你是波士顿咨询集团（Boston Consulting Group, BCG）的顶级顾问，挑剔并批判我的方案，告诉我如何能写得更好。”AI 随即指出了关键疏漏，例如需要补充财务影响预测以及更明确的汇报层级。Steven 吸收了这些反馈，并逐一优化了自己的备忘录。 ```mermaid flowchart TD accTitle: 高阶 AI 交互与低阶偷懒的认知路径对比 accDescr: 对比展示深度独立思考后使用 AI 批判，与直接让 AI 生成内容的脑力锻炼差异。 subgraph HighLevel ["深度心智锻炼路径（CEO 式工作法）"] A["独立思考与草拟方案<br/>耗时 4 小时深度推演"] --> B["多模型多角度提问<br/>设定顶级顾问人设求批判"] B --> C["吸收关键漏洞反馈<br/>完善财务预测与架构"] C --> D["强化基底核回路<br/>加深认知与业务掌控力"] end subgraph LowLevel ["认知退化快捷路径（被动偷懒法）"] E["输入简短指令<br/>30 秒生成岗位描述"] --> F["全盘接受 AI 产出<br/>放弃推导与思考过程"] F --> G["大脑练习剥夺<br/>核心认知能力萎缩"] end ``` 这种交互展现了健康的人机协作模式，但 Steven 直言大多数人根本不会这么做。凭借 ChatGPT 的记忆功能，他完全可以在 30 秒内让 AI 直接代笔，省下三个半小时。而他之所以强迫自己选择费力耗时的艰难路径，是因为他深知 23 岁时高强度写作与表达对他人生的塑造意义——如果年轻时没有坚持整整 5 年每天写作、将复杂思想浓缩在推特（Twitter）的 140 字限制内，他绝不可能拥有如今的商业洞察力和底层认知。Sejnowski 博士对此赞叹：“这就是为什么你是公司的 CEO，而别人不是。” ### [44:02](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2642s) 基底核与机械式练习的重要性 Sejnowski 博士从脑科学角度解释了 Steven 的成功秘诀：这种长期的反复打磨调动并强化了大脑的**基底核（Basal Ganglia）**。基底核掌管习惯形成与程序性动作，其运作极度依赖“重复、练习、再练习”。只有当基础练习深植于基底核中形成自动化回路，高级认知功能才能在此基础上游刃有余地发挥。 人类的学习依赖两大认知系统的协同运作，而当前的现代教育却走入了致命误区。学校教育逐渐抛弃了所谓的“机械式学习”（Rote Learning），甚至将“死记硬背”贬低为落后产物。Sejnowski 博士直言这种观念大错特错： - **流利度（Fluency）的建立**：无论是数学计算还是阅读理解，掌握一门学科的前提是通过大量记忆与重复练习建立底层直觉。 - **AI 架构的天然短板**：现代大语言模型根本没有模拟人类基底核的运作结构，它们只是概率生成，缺乏通过重复实体经验沉淀出的内化神经回路。 - **基础缺失的认知代价**：抛弃了基底核的刻意练习，学生就失去了认知飞跃的地基，最终既无法形成深层记忆，也难以建立真正的思维敏捷度。 ### [44:50](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2690s) 拼写退化与自然拼读法的缺失 AI 工具对人脑最立竿见影的侵蚀体现在语言文字技能的退化上。Steven 坦言自己现在打字拼写越来越懒散，因为大语言模型比传统的文字处理软件（如 Word 拼写检查）聪明得多。传统工具只能纠错，而大模型能完全“洞察”你的意图。现在输入一个 12 个字母的复杂长词，哪怕只敲对 6 个字母，AI 也能准确无误地理解。 为了向两位专家直观展示这一点，Steven 现场闭上双眼，在 iPad 上盲打了一句惨不忍睹、几乎每个单词都拼错的句子：“Tell me everything I know about Daniel Amen”。即便文本被严重肢解，ChatGPT 依然毫发无损地识别出来，并给出了完整的 Daniel Amen 博士生平档案。Steven 感叹，这种便利性会反向驯化人类大脑，让人产生“以后再也不需要学习拼写，只要拼个大概即可”的依赖心理。 Daniel Amen 博士对此深有同感并深表忧虑。他们这代人是通过**自然拼读法（Phonics）**学习单词发音与拼写的，但加利福尼亚州的学校系统在整整一代人的时间里彻底废弃了自然拼读教学。这种教育方针的巨变重塑了孩子的大脑结构，直接导致整整一代年轻人失去了扎实的拼写与语言解构能力。这再次印证了脱离重复性、程序化记忆训练（做题、反复演算直至形成无意识反射）对大脑发育带来的系统性破坏。 ### [46:36](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2796s) 间隔效应与大脑真正的学习机制 面对充斥海量信息但难以留存的困境，Steven 请教 Sejnowski 博士：基于一个多世纪以来的大脑科学研究，究竟如何才能学得更好、记得更牢？Sejnowski 博士指出，脑科学界早在一百年前就揭示了长期记忆的核心铁律——**间隔效应（Spacing Effect）**，这一理论最早可追溯至德国心理学家赫尔曼·艾宾浩斯（Hermann Ebbinghaus）。 - **告别突击学习**：若要长久掌握某项知识，绝对不要试图在一次马拉松式的通宵突击中完成所有的学习负荷。 - **规律的时间间隔**：必须将学习时间分散，学完新知识后，在第二天进行回想复习，隔一周后再次提取巩固。 - **大脑固化机制**：时间的间隔给大脑神经突触留出了结构重组的生理空隙，使瞬时记忆真正向长期记忆凝固。 令人讽刺的是，学校往往要求学生死记硬背以应付考试，却从未教授过这种最基本的科学学习法。Amen 博士补充道，学校教育的首要任务本该是“教学生如何热爱并呵护自己的大脑，然后再教他们如何科学地学习”，现实却恰恰本末倒置。 ### [48:16](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2896s) 克服拖延与睡眠中的记忆巩固 针对现代人普遍的拖延症，Sejnowski 博士结合他与 Barbara Oakley 教授共同打造的全球现象级公开课《学会如何学习》（*Learning How to Learn*，全球学员已超 600 万人）给出了经过神经科学验证的解决方案： 拖延的本质是大脑面对困难任务时感受到的心理阻滞（Mental Block）或能量壁垒。大脑的痛觉中心在预想困难时被激活，驱使个体转向轻松的事物。强行硬碰硬往往收效甚微，打破僵局的有效策略如下： 1. **设立微小的初始目标**：对自己承诺今天只投入 20 分钟启动任务，即使只是梳理一下思路或列个大纲。 2. **设定倒计时闹钟**：在这 20 分钟内全神贯注，闹钟一响立刻停下，绝不恋战。 3. **利用睡眠离线加工**：在这短短 20 分钟建立心智线索后，潜意识会在你入睡期间持续对这些线索进行后台运算与重组。 4. **第二天循序渐进**：次日醒来再投入下一个 20 分钟，此时前一天的神经链接已初步建立，任务壁垒显著降低。 民间常说的“睡一觉再做决定（Sleep on it）”有着深刻的神经科学依据。Sejnowski 博士解释，睡眠期间大脑正在进行高阶的**记忆巩固（Memory Consolidation）**：它将白天的全新经历小心翼翼地缝合进原有的长期记忆网络中，确保二者不产生逻辑冲突，同时剔除白天的琐碎干扰，提炼核心规律。这就是为什么清晨醒来时，前一晚看似杂乱无章的难题往往会变得豁然开朗。 在本段讨论的尾声，Steven 反思了自己的播客准备习惯：面对长达 20 页的研究资料，他常常以为只要一遍又一遍反复重读就能记住所有细节。Sejnowski 博士当即纠正道，单纯的“机械重读”是大脑制造的虚假学习幻觉，学生们遇到理解瓶颈时，如果只是一味地拿头硬撞墙壁般重复阅读，不仅无法真正掌握，反而会彻底封死认知重构的可能。 ### [50:39](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3039s) 潜意识整合与“滑雪会议”：为什么大脑需要间歇 当你在思考问题撞上思维之墙、无论如何也搞不懂时，最正确的做法绝不是继续硬撑，而是立即站起来走动，去散步、做饭或打理花园。大脑会在高强度输入下极快地达到饱和，把事情暂时放下，才能让潜意识接管并继续加工。许多人误以为在密集的会议或冗长的学术讲座之间安排休息是浪费时间，但事实上，这些中场休息才是一整串演讲中最不可或缺的部分，它给大脑留出了沉淀和反刍的空间。 特里·塞诺夫斯基（Dr. Terry Sejnowski）分享了他最推崇的会议形式——“滑雪会议”（Ski Meeting）。这种会议通常选在滑雪胜地举行，有着独特的日程节奏：上午先听两小时的高浓度学术报告；随后大家出发去滑雪，在这段身体运动的时间里，大脑潜意识其实在飞速消化上午接收的信息；到了傍晚归来，大脑已经焕然一新，此时再进行两小时的讲座，参会者便能极其高效地吸收新知并融会贯通；夜间入睡后，大脑会进一步进行记忆巩固。特里将这个过程比喻为“揉面团”，必须来来回回反复揉压，这远比把所有演讲塞进几天、等会议彻底结束后再去消化要高效得多。 ```mermaid flowchart TD accTitle: 滑雪会议学习与整合模型 accDescr: 展现上午输入、运动反刍、傍晚再输入与睡眠巩固的高效循环 subgraph Phase1["白天学习与潜意识处理"] A["上午：2小时学术讲座<br/>（高强度信息输入）"] --> B["白天：户外滑雪运动<br/>（放空显性思维，潜意识整合反刍）"] end subgraph Phase2["二次吸收与睡眠巩固"] B --> C["傍晚：2小时讲座<br/>（大脑恢复活力，高效融会贯通）"] C --> D["夜间：深度睡眠<br/>（如反复揉面，长时记忆沉淀）"] end ``` --- ### [51:56](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3116s) 商业赞助插播：Stan Store 与沃达丰 Business.connected 史蒂文·巴特利特（Steven Bartlett）在此处插入了两则品牌赞助： - **Stan Store**：史蒂文作为联合创始人的平台，主打一键将创作者的知识、技能与经验变现，无需编程背景即可销售数字产品、咨询服务和社群订阅，目前入驻创作者已累计创造3亿美元收益。 - **沃达丰 Business.connected**：面向初创团队与中小微企业的数字化扶持计划，提供覆盖人工智能、数字化营销、电商运营等实战技能的免费教学内容与创始人经验分享系列。 --- ### [53:54](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3234s) 优生与早期养育：如何为下一代构建“大脑储备” 抛开人工智能不谈，史蒂文提到自己计划在未来几年内成为父亲，并询问如何从源头上确保孩子的脑健康。丹尼尔·亚蒙（Dr. Daniel Amen）指出，保护孩子的大脑必须从“受孕前”就开始。他引入了“大脑储备”（Brain Reserve）的核心概念——即大脑拥有的额外组织与缓冲功能，用以抵御未来可能降临的各种压力与损伤。大脑储备的奠基始于优质的精子与卵子，因此准父母在备孕阶段就应尽可能过上干净、健康的生活。 一旦妻子怀孕，男方最重要的责任就是避免让她承受过大压力。胚胎的大脑在受孕后大约第21天就已经开始发育，往往比孕妇意识到自己怀孕的时间还要早。当婴儿出生后，他们最需要的并非昂贵的早教玩具，而是父母全心全意的陪伴、目光接触（Eye Contact）、拥抱抚触以及轻声歌唱。 特里补充指出，除了身体抚触，“早期语言环境”的密度对智力发展至关重要。婴儿虽然要到大约18个月大才开口说话，但父母所说的每一个字都在持续刺激其大脑回路。研究表明，在缺乏日常语言交流的家庭中长大的孩子，未来的学业表现明显较差，这在部分贫困家庭中尤为突出，因此尽早且丰富地向婴儿灌输语言是一项关键投资。 - **父母的身教范式**：父母每一天的所作所为、言谈举止，以及父亲对待母亲的方式，都在无形中向孩子示范“何为健康”或“何为病态”。 - **陪伴的本质是实体时间与倾听**：亚蒙在《养育心理强大的孩子》（*Raising Mentally Strong Kids*）中强调，建立纽带的核心在于投入真实的物理时间和耐心的倾听。 - **屏幕对亲子关系的侵蚀**：现代科技让父母深陷手机屏幕，餐桌上全家各看各的设备，失去了相互凝视与沟通，甚至导致孩子转而向能耐心倾听且不打断的 AI 寻求心理反馈。 --- ### [57:41](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3461s) 心理脆弱的一代与超越性信仰的神经学机制 面对史蒂文关于“现代过度庇护的育儿文化是否在造就心理脆弱的孩子”这一提问，亚蒙坦言：“当代年轻一代陷入了历史上最严重的困境。”这种困局是由劣质饮食、电子产品依赖、负面新闻轰炸以及极端的社会撕裂共同酿成的，使年轻人长期处于皮质醇超标的慢性应激状态。更荒谬的是政治立场带来的彻底孤立——仅仅因为投票倾向不同就与人绝交，在孤独感本就蔓延的时代人为切断了与半数人群的连接。 关于宗教信仰与超验体验对大脑的影响，两位专家给出了颠覆直觉的神经生物学解释： ```mermaid flowchart LR accTitle: 信仰对心理状态与大脑区域的影响 accDescr: 展现有无信仰在抑郁风险、颞叶结构及基底核激活上的神经学表现 subgraph Mindset["心理存在感"] M1["虚无主义 / 随机概率<br/>（抑郁风险增加3倍）"] M2["超验信仰 / 使命感<br/>（更长寿、更高幸福感）"] end subgraph BrainMechanics["脑神经机制"] B1["颞叶（上帝区域）<br/>信徒拥有更大体积；异常放电诱发超验神圣感"] B2["说方言 / 灵性通灵<br/>额叶活动抑制（自我控制放低） + 基底核飙升（多巴胺极度活跃）"] end M2 -.-> B1 M2 -.-> B2 ``` - **缺乏信仰与抑郁风险**：若认定生命只是随机的偶然产物且毫无意义，会诱发存在主义的虚无感，这类人群患抑郁症的风险比坚信存在更高力量的人高出整整3倍；怀揣使命感（Purposeful）的人往往寿命更长、心理更健康。 - **颞叶与“上帝区域”**：神经影像研究发现，信徒的大脑颞叶（位于太阳穴下方、眼眶后部）体积反而比非信徒更大。该区域被称为大脑的“上帝区域”（God Area）；当颞叶出现癫痫样放电时，个体会体验到置身于神圣存在之中的超验感（如保罗在去往大马士革路上突发异象的推测）。 - **神盔实验（God Helmet）**：加拿大劳伦森大学的迈克尔·珀辛格（Michael Persinger）通过刺激受试者的右侧颞叶外侧，成功在实验室环境下诱发出强烈的“被感知存在感”（Sensed Presence），使受试者真切感受到房间里有神圣实体降临。 - **祈祷与“说方言”的脑扫描实验**：亚蒙对宗教仪式中的“说方言”（Speaking in Tongues）进行了专门的脑成像研究。结果显示，60%的受试者额叶活跃度显著下降（表明有意识的自我审查与逻辑控制被主动关闭），同时基底核（Basal Ganglia）的兴奋度出现爆发式飙升——其刺激模式与大脑受到可卡因冲击时的反应高度相似。 ### [01:01:35](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3695s) 打造健脑国家的单一准则与最强“药物”：运动 当史蒂文·巴特利特询问，如果可以通过签署行政命令来打造一个“大脑健康国家”，会颁布什么政策时，丹尼尔·阿门博士（Dr. Daniel Amen）给出了一个出人意料却极其有力的答案：只需要所有政府部门在做决策前都回答一个核心问题——“我们正在做的事情，对国民的大脑是有益还是有害？”阿门博士感慨自己从事脑科学研究数十载，始终坚信只要人们能带着对自我、家庭和国家的爱，基于科学信息去诚实回答这个问题，社会的整体健康就会发生质的改变。 特里·塞诺夫斯基博士（Dr. Terry Sejnowski）随即指出，对于大脑乃至整个身体而言，“迄今为止最好的药物就是运动”。运动能够高效泵送血液，为大脑输送大量的氧气与营养素，同时强化心脏功能与免疫系统。这种运动并不苛求成为专业运动员，尤其对老年人而言，规律的散步就是极佳的脑部锻炼；然而现代的孩子们沉迷于各种电子屏幕，普遍严重缺乏身体活动。阿门博士补充道，在他提出的旨在预防和逆转脑衰退的“BRIGHT MINDS”十一大核心风险因素模型中，运动几乎是唯一能同时攻克所有指标的全能方案。 - **B（Blood flow，血流量）**：运动能直接增加脑部血液循环，改善供血不足。 - **R（Retirement & Aging，退休与衰老）**：积极运动能有效降低生物学年龄。 - **I（Inflammation，炎症）**：运动具备强大的全身性抗炎作用。 - **G（Genetics，基因表达）**：促进健康基因的表达与开启。 - **H（Head trauma，脑外伤预防）**：维持散步习惯能提升平衡感，显著减少老年人因跌倒导致的脑外伤风险。 - **T（Toxins，排毒）**：运动排汗是身体排出代谢毒素的自然路径。 - **M（Mental health，心理健康）**：运动不仅能促进多巴胺释放提升动力，还能增加血清素分泌，达成神经递质的完美平衡。 --- ### [01:03:38](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3818s) 呼吸与咀嚼：调节副交感神经与海马体活力的生理开关 讨论延伸至日常呼吸对脑健康的深远影响。阿门博士指出，人们完全可以通过一套特殊的呼吸节律，几乎瞬间提升心率变异性（HRV，心脏健康与自主神经调节的关键指标），从而迅速平复情绪。他将其命名为“15秒呼吸法”：吸气 4 秒，屏息 1.5 秒，呼气 8 秒，再屏息 1.5 秒。这套呼吸法的生理学核心在于“呼气时间两倍于吸气时间”，它能强力激活迷走神经、提高副交感神经张力（Parasympathetic tone）。阿门博士直言，当恐慌发作时，服用赞安诺（Xanax）等抗焦虑药虽然见效快，但长期会带来一系列认知和成瘾隐患，而通过以腹部起伏为主的腹式深呼吸，就能达到天然、安全且快速的镇静效果。 ```mermaid flowchart LR accTitle: 15秒腹式呼吸循环 accDescr: 吸气4秒、停顿1.5秒、呼气8秒、停顿1.5秒的副交感神经激活路径 A["腹式吸气 (4秒)<br/>腹部向外隆起"] --> B["短暂停顿 (1.5秒)<br/>闭气蓄力"] B --> C["缓慢呼气 (8秒)<br/>激活副交感神经"] C --> D["短暂停顿 (1.5秒)<br/>彻底放松"] D -->|进入下一循环| A ``` 除了呼吸，机械性的“咀嚼”动作也对脑部有着意想不到的保护作用。文献证据表明，咀嚼能够直接刺激海马体区域的神经元活性，延缓认知退化；而在动物实验中，剥夺咀嚼功能已被证实会导致学习记忆能力的明显受损。阿门博士指出，现代快餐工业为了让消费者吃得更快、吞咽得更快，刻意滤除和去掉了大部分膳食纤维，这种高度精细化的加工食物正在悄悄剥夺人们咀嚼的机会，进而潜移默化地削弱海马体功能。 --- ### [01:04:58](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3898s) GPS 与表观遗传学：导航依赖与小鼠气味恐惧的世代传承 过度依赖外在技术工具对大脑结构的重塑同样令人担忧，其中 GPS 导航软件的滥用尤为典型。长期将空间记忆和寻路功能外包给手机，会导致负责空间定位和记忆的海马体缺乏锻炼，长此以往可能引起海马体萎缩。阿门博士分享了一个令人深思的临床现象：早年他在刚做精神科医生时，如果一位在某座城市居住了 30 年的老人突然迷路，家属会非常警惕并立刻求医，这通常是阿尔茨海默病的早期信号；然而如今，迷路的老人只需对着手机说一句“Siri，带我回家”，病情就被现代科技严密地遮蔽了，导致确诊时间被大幅推迟。 这一现象引发了关于表观遗传学（Epigenetics）的讨论：如果当代人完全依赖手机导航而不去记忆地图，这种生活方式是否会通过表观遗传标记遗传给下一代？特里·塞诺夫斯基博士对此持审慎态度，认为这通常需要经由长期的生理机制传导（例如孕期母体的慢性压力激素），但阿门博士随之引用了一项经典的跨代遗传实验： - **樱花气味电击实验**：科研人员将雄性小鼠暴露在樱花气味（苯乙酮）中并施加轻微电击，使小鼠对该气味建立强烈的恐惧条件反射。 - **跨代传递现象**：令人震撼的是，这些小鼠从未接触过电击的子代（第二代），乃至孙代（第三代），在首次闻到樱花香气时，均自发表现出剧烈的恐惧反应。 - **演化生物学解释**：塞诺夫斯基博士指出，嗅觉系统非常独特，其神经信号直接投射至海马体和边缘系统。从演化角度来看，若父辈能将环境中“致命毒物或危险捕食者的特征气味”刻入表观遗传并传递给后代，后代无需付出生命代价进行试错即可规避风险，这在生存演化上具有极高的适应优势。 --- ### [01:07:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4028s) 人工甜味剂的隐形代价与肠道微生态 对于大众普遍认为“无害且无热量负担”的人工甜味剂（代糖），阿门博士给出了严厉警告：“它们并不好，也绝非免费的午餐。”他分享了自己的亲身经历：他年轻时曾将无糖无卡汽水当成最好的饮品狂喝不止，结果在 35 岁时患上了关节炎。后来一名患者告诉他，戒除阿斯巴甜后关节炎自愈了；阿门博士抱着怀疑的态度戒掉了代糖汽水，关节炎竟真的彻底消失，而当他重新尝试饮用时，关节疼痛立刻复发，这让他彻底认清了代糖的负面影响。 人工甜味剂对神经和身体的核心破坏途径之一是破坏肠道微生态。人体肠道内寄宿着多达 100 万亿个微生物，它们负责消化食物并合成大量关键的神经递质。研究已经发现，常见的三氯蔗糖（Sucralose，如 Splenda）等代糖会显著破坏肠道菌群的多样性，降低益生菌数量，进而通过“肠-脑轴”损害大脑认知与情绪调控。阿门博士指出，阿斯巴甜被引入食品供应链已有数十年，现代年轻人面临的焦虑和心理危机，可能并不仅仅是社交媒体的单一结果，长期摄入代糖引起的菌群与表观遗传改变同样是不可忽视的叠加因素。他建议大家在日常饮食中严守“是否益脑”的原则，例如多吃西兰花；如果要吃汉堡，可以去掉面包和精加工酱料，选择草饲牛肉搭配新鲜沙拉。 --- ### [01:08:50](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4130s) 慢性环境噪音与认知负荷：大脑未意识到的皮质醇压力 慢性背景噪音对大脑的隐性侵蚀同样常被低估。阿门博士坦言，他过去的家距离高速公路仅有三栋房子的距离，访客每次拜访都会抱怨车流轰鸣刺耳，但他自己却完全听不见，因为大脑早已学会了自动将噪音过滤掉（他打趣道，家里有五个姐妹的环境也锻炼了这种“屏蔽噪音”的能力）。 然而，塞诺夫斯基博士与史蒂文严肃指出，**“主观上适应了噪音”并不等同于“身体免受了伤害”**： - **潜意识生理应激**：即便意识层面忽略了轰鸣声，大脑皮层与听觉中枢依然在后台持续接收并处理声音信号，这种特殊的适应状态耗费了大量的神经计算资源。 - **皮质醇暗中升高**：长期的环境低频轰鸣（如交通干道、城市背景噪音）会引起下丘脑-垂体-肾上腺轴（HPA轴）的微弱持续激活，使皮质醇水平在不知不觉中悄然抬升。 - **认知功能受损**：持续处于噪音环境中的人群，其工作记忆能力和注意力调节机制会受到明显损伤，这种认知耗损在神经系统尚在发育的儿童以及生理储备下降的老年人身上表现得尤为剧烈。 --- ### [01:10:03](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4203s) 幕后社群招募与多屏幕任务处理的隐患 在访谈间歇，史蒂文插入了一段简短的公告，宣布启动名为“DOAC Circle”的专属私密社群。该社群首批仅开放 10,000 个席位，加入的成员将拥有史蒂文的直接互动通道，并能获取节目录制时使用的 iPad 备课提纲、幕后花絮以及从未公开发布过的嘉宾对谈内容与删减片段，观众可通过 DOACcircle.com 申请加入。 重回对话后，史蒂文将焦点转向了现代人极其普遍的数字生活常态——跨屏多任务处理（Multitasking across multiple screens）。现代人的注意力被高度碎片化割裂：眼前开着电视，手里刷着智能手机，膝盖上放着 iPad，桌上还立着电脑屏幕。面对这种前所未有的信息过载与多任务并行模式，神经科学研究正在揭示其对人类专注力、大脑回路以及注意力缺陷多动障碍（ADHD）激增之间令人不安的深层关联。 ### [01:11:02](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4262s) 多任务处理、脑岛与抑郁症的本质 多任务处理（Multitasking）正在悄然改变人类的大脑结构，它会训练大脑变得极易分心，并导致前扣带回（Anterior Cingulate Cortex）、脑岛（Insula）以及内侧前额叶皮层（Medial Prefrontal Cortex）的灰质密度显著下降。Dr. Daniel Amen 借此提到了他即将发表的一项关于“希望”的重磅研究：团队对 7,500 名患者进行了希望量表测试，评估他们是否相信自己有能力让明天变得更好。结果显示，那些缺乏希望的人群整体前额叶功能明显低下，其中脑岛的活动抑制尤为剧烈，成为该组数据中统计学意义最显著的信号。脑岛在神经学中常被称为连接身体感受与情绪认知的“神经岛屿”。 Dr. Terry Sejnowski 补充道，在抑郁症患者中同样能观察到前扣带回区域的异常低活性；目前神经医学界甚至利用深度脑刺激技术（Deep Brain Stimulation, DBS）对该区域进行电刺激，以此成功改善了部分难治性抑郁症患者的症状。然而，Dr. Daniel Amen 强调临床上绝不能将抑郁症视为单一维度的疾病。正如医生不能将“胸痛”当作最终诊断一样——胸痛的诱因可能是心肌梗死、心律失常、心脏感染、胃肠胀气甚至剧烈的悲伤；抑郁症亦然。从脑部单光子发射计算机断层成像（SPECT）来看，有些抑郁症患者额叶过度活跃，有些则活性不足，还有些表现为边缘系统的过度亢奋。他在其著作《治愈焦虑与抑郁》（*Healing Anxiety and Depression*）中，通过脑成像技术将这种心理综合征细分为了 7 种完全不同的脑功能生理亚型。 - 7,500 名患者的临床研究显示，“希望感”缺失与脑岛活动的大幅减弱存在极强正相关。 - 抑郁症并非单一疾病，其脑影像学表现各异，部分患者可通过前扣带回的深度脑刺激获得缓解。 --- ### [01:12:43](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4363s) ADHD 的激增原因与药物治疗的真相 当被问及近年来注意力缺陷多动障碍（ADHD）确诊人数飙升究竟是基因使然还是现代生活方式的产物时，Dr. Daniel Amen 明确指出这是“双重作用”的结果。ADHD 具有极其明显的家族遗传性，临床接诊多动、焦躁、冲动、缺乏条理且习惯拖延的患儿时，父母身上往往能找到完全一致的行为模式。但与此同时，脑外伤同样能够引发 ADHD，特别是撞击额叶的损伤——这也是为什么绝不应允许儿童用前额去顶足球的原因；此外，现代数字环境带来的持续过载信息输入，也在不可逆地剥夺大众的专注力。 谈及药物干预时，人们往往对利他林（Ritalin）等中枢神经兴奋剂谈虎色变，甚至将其直接等同于“安非他命”。然而 Dr. Daniel Amen 引用了一项最新研究：服用药物治疗的 ADHD 患儿，其前额叶皮层的发育体积反而大于那些患病却未接受药物治疗的患儿。他指出，社会往往在妖魔化药物，却忽视了不治疗带来的残酷代价——确诊 ADHD 的人群中有三分之一无法完成高中学业。治疗药物确实存在食欲下降或影响睡眠等副作用，但任由大脑功能失调恶化的隐性代价更为沉重。虽然他本人创办了营养补充剂公司并极力推崇脑部神经营养优化与脑电生物反馈疗法（Neurofeedback），但他坚定主张，当非药物干预无效时绝不能排斥正规用药：“对一个真正患有 ADHD 的孩子隐瞒药物，就像剥夺近视患者佩戴眼镜的权利一样。” 回溯历史，ADHD 并非现代才凭空出现的产物，早在 1910 年的医学文献以及第一版《精神疾病诊断与统计手册》（DSM）中就有记载，当时被称为“轻微脑功能障碍”。令人深思的是发生率的变化：老一辈上学时，一个班级往往只有一两个注意力异常的孩子，而如今每个班级常常能查出八到十个。这种急剧增加与高度加工食品的大量摄入、电子屏幕泛滥、父母自身的注意力涣散以及缺乏耐心的教学模式密不可分。 ```mermaid flowchart TD accTitle: ADHD成因与干预多维分析 accDescr: 图表展示了ADHD的内外部诱因、不干预的潜在后果以及综合治疗途径。 subgraph Causes["ADHD 诱因构成"] Genetics["遗传基因因素<br/>（家族聚集性）"] Physical["脑部物理损伤<br/>（前额顶球等额叶外伤）"] Lifestyle["现代生活环境<br/>（超加工食品/屏幕过载/分心环境）"] end subgraph Impacts["未经干预风险"] FrontalDecline["前额叶皮层发育受阻"] DropoutRate["严重学业障碍<br/>（1/3 未能完成高中学业）"] end subgraph Solutions["科学干预手段"] Medication["规范药物治疗<br/>（如 Ritalin 保护前额叶发育）"] LifestyleRx["辅助治疗方案<br/>（脑部营养/生物反馈训练）"] end Genetics --> Impacts Physical --> Impacts Lifestyle --> Impacts Impacts --> Solutions ``` --- ### [01:15:50](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4550s) 消极思维对脑结构的破坏与性别差异 在谈到近期最令人振奋的研究时，Dr. Daniel Amen 介绍了他主导的一项关于“消极偏见与大脑结构”的专题研究。通过“正负面偏见问卷”对比发现，长期沉溺于消极负面思维的人，其前额叶皮层的功能活动会受到极大的抑制。他指出，盲目乐观不可取，人脑需要保留约 15% 的谨慎与批判性思维以规避风险；但如果陷入长期的慢性消极情绪，大脑生理机能将遭到严重实质损伤。 消极心态与神经退行性疾病之间存在着令人震惊的性别差异。临床数据显示，如果一名女性患有抑郁症，其罹患阿尔茨海默病（Alzheimer's Disease）的风险将直接翻倍；而如果一名男性患有抑郁症，其阿尔茨海默病的患病风险更是会飙升整整四倍。Dr. Terry Sejnowski 结合公共卫生数据补充道，在新冠疫情期间，女性的抑郁发生率暴增了一倍，而男性在数据上并未出现这种猛增；更残酷的是，当社会生活恢复正常、学生重返校园后，男性的心理状态逐渐回稳，女性的抑郁比例却始终顽固地停留在疫情高位。 Dr. Daniel Amen 对这一神经生物学现象给出了原理解释： 1. **神经递质基线差异**：既往研究表明，女性大脑中的血清素（Serotonin）合成水平比男性平均低 52%，这导致女性天然更容易受到情绪低谷的侵袭，整体抑郁患病几率本就是男性的两倍。 2. **情感与连接中枢脆弱性**：女性拥有体积更大的边缘系统（Limbic System），该区域负责情感共鸣与人际纽带的建立，在社会隔离与情感冲击面前表现出更高的生理易感性。 3. **感染引发脑神经炎症**：新冠病毒感染会直接在人脑的边缘系统诱发明显的神经性炎症。Dr. Daniel Amen 对比了患者感染前后的脑部扫描影像，清晰呈现出病毒感染对边缘结构造成的剧烈炎症风暴。 --- ### [01:18:14](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4694s) 维护终身脑健康的三大基石与桑拿疗法 当主持人请两位专家为大众开出守护大脑、对抗认知衰退的终极处方时，Dr. Daniel Amen 与 Dr. Terry Sejnowski 提炼出了延缓衰老、活到百岁依然保持敏锐思维的核心行动法则： 第一是坚持体育锻炼（Exercise）。运动不仅能强健体魄，更是维持海马体与前额叶血液循环的生理基础。 第二是启动早晨的“认知定向”。每天清晨醒来时，主动向大脑发出指令告诉自己“今天将会是极好的一天”，刻意强迫前额叶皮层去搜寻生活中的正确之事，而非陷入对错误与缺陷的习惯性反刍，以乐观与感恩重构神经网络。 第三是足量补充 Omega-3 脂肪酸。大脑细胞膜干重的 25% 是由 Omega-3 构成的，它具有不可替代的抗炎与修复效能。然而在现代饮食结构下，绝大多数人群体内的 Omega-3 水平都处于极其危险的匮乏状态，必须通过多吃深海鱼或服用高品质纯化补充剂来进行弥补。 此外，两人特别提到了“边运动边学习”的神奇增效机制。如果在有氧运动的同时学习新知识，激增的脑血流会精准灌注至海马体，使记忆牢固度和神经重塑效率产生飞跃。主持人也分享了一名科学家的经历：该学者发现自己在桑拿房里复习考试内容并在考前利用这种环境能够显著提升成绩。Dr. Daniel Amen 对此表示赞同，并指出《美国医学会杂志·精神病学》（*JAMA Psychiatry*）曾刊登研究证实，单次桑拿浴就能显著减轻患者的抑郁症状；长期坚持蒸桑拿的人群，其最终罹患阿尔茨海默病的风险是全人群中最低的，这得益于高温诱导的血管扩张对大脑血液稳态的积极重塑。 --- ### [01:20:51](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4851s) 终极寄语：睡眠对记忆巩固的不可替代性 在全场讨论走向核心升华之际，面对关于人工智能时代神经科学与大脑健康的最终拷问，Dr. Terry Sejnowski 毫无保留地给出了全人类最应守护的大脑防御机制——**睡眠（Sleep）**。 Dr. Terry Sejnowski 强调，睡眠绝对不仅是身体机能从疲劳中被动恢复的时间，更是人类记忆进行主动巩固（Consolidation）的决定性窗口。日间所经历的庞杂信息、学到的技能与产生的情感体验，只有在深度睡眠的神经震荡中，通过海马体与大脑皮层之间精密复杂的双向信号交互，才能被筛选、编码并最终整合到深层的大脑皮层回路之中，转化成稳定、持久的长时记忆。失去睡眠的保护，一切大脑储备与认知健康都将无从谈起。 ### [01:21:05](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4865s) 睡眠、运动与放任的“特洛伊木马” Dr. Terry Sejnowski 解释道，睡眠对大脑皮层整合情景记忆（episodic memories）至关重要。令人痛心的是，如今的孩子们面临着极其激烈的大学升学竞争，往往最先被牺牲的就是睡眠，而这恰恰发生在他们大脑发育最关键的黄金时期。Dr. Terry 明确强调，睡眠与运动是大脑健康的基石，重要性无出其右。 当主持人 Steven 询问 Dr. Daniel Amen 对整场讨论的结语时，Dr. Amen 用了一个生动的比喻：人类如今就像是猛然敞开了谷仓大门，任由一匹脱缰野马狂奔进学校、企业和千家万户。然而，在放任它闯入之前，我们甚至没有停下来思考过——这究竟是一份馈赠，还是一匹企图掏空我们的“特洛伊木马”。人类总是“在理解后果之前就全盘拥抱了便利”，正如过去在电子游戏、智能手机、社交媒体、大麻、酒精、阿片类药物、高果糖玉米糖浆和阿斯巴甜上重蹈覆辙一样。Dr. Amen 警告，我们必须用智慧驯服这匹狂奔的野马，否则它将无情地踩踏下一代。评判这一切的核心标准唯有一条：“这究竟对我的大脑有益，还是有害？对我们的集体大脑是有益还是有害？”并用爱与科学去诚实作答。 - **睡眠与记忆**：大脑皮层依赖充足睡眠来稳固情景记忆，青少年期缺觉对大脑发育损伤极大。 - **健康核心基石**：Dr. Terry 认为保持大脑功能最关键的两件事就是睡眠与规律运动。 - **便利与后果的脱节**：科技产品往往先以“极度便利”普及，后才暴露出长期的认知与生理代价。 --- ### [01:22:54](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4974s) 新技术的历史周期与“走困难的路” 主持人 Steven 坦言这番对话让他比刚进录音室时感到更加焦虑，但他深刻认同关于历史周期的分析。回顾精制加工食品、社交媒体乃至约会软件的普及，几乎都遵循着同一条发展曲线： ```mermaid flowchart LR accTitle: 技术与生活方式演进的三阶段周期 accDescr: 展示新技术从被利益驱动者粉饰、到后果显现并被研究、最终推动个体改变行为的三阶段模型。 subgraph Phase1["第一阶段：狂热与掩盖"] A["新产品/新技术问世"] --> B["利益相关者粉饰太平<br/>声称毫无风险"] end subgraph Phase2["第二阶段：反思与暴露"] B --> C["负面后果逐渐显现"] C --> D["科学研究介入<br/>发现隐形代价(Trade-off)"] end subgraph Phase3["第三阶段：觉醒与应对"] D --> E["公众行为觉醒与调整<br/>学会主动节制"] end ``` 意识到所有技术红利背后都有隐形代价（Trade-off）后，Steven 开始反思 AI 的代价。AI 带来的短期生产力暴涨显而易见，但潜在代价是批判性思维的退化、独立思考能力的丧失，甚至是由于沉迷虚拟伴侣（比如逼真的 AI 角色）而损害现实的人际关系。 Steven 认为，面对未来十年，真正的竞争护城河与对冲策略或许是“逆向而行”——当所有人都在向右奔向即时生成的便利时，选择向左走，坚持“走困难的路”（doing it the hard way）。正如棉花糖实验中展示的延迟满足一样，在关乎核心思维和深度表达时主动克制不用 AI：当思考至关重要时，坚持自己独立思考；当沟通至关重要时，坚持亲笔真诚表达。Dr. Amen 对此表示赞同，强调父母必须以身作则，因为“每天你的行为都在向孩子示范什么是健康，什么是不健康”。 - **对冲 AI 的核心策略**：逆向克制，不将核心认知劳动全盘外包给算法。 - **父母的榜样力量**：言传不如身教，父母对待科技和健康的生活模式会直接被下一代复刻。 --- ### [01:25:48](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=5148s) 终极提问：面对下一个健康挑战 在节目推荐了两位专家的著作——Dr. Daniel Amen 的《Raising Mentally Strong Kids》《Change Your Brain Every Day》以及 Dr. Terry Sejnowski 的《The Deep Learning Revolution》《ChatGPT and the Future of AI》之后，迎来了节目的固定闭幕传统：由上一位未公开身份的嘉宾留下一个问题。 本期留下的问题是：“你是否做好了察觉下一个健康挑战的准备？你能否敏锐注意到它的发作？即使意味着生活方式或生存状态的重大改变，你又将如何应对？” 面对这个问题，两位专家给出了风格迥异但极其真实的回答： | 嘉宾 | 是否做好准备 | 应对策略与人生哲学 | |:--- |:--- |:--- | | **Dr. Daniel Amen** | **是** | 目标高度清晰，坚决要保持健康活力、远离阿尔茨海默病。只要为了达成目标，无论需要做出多大的生活方式调整，他都会毫不犹豫地“全力以赴（all in）”。 | | **Dr. Terry Sejnowski** | **可能没有** | 承认随着年龄增长，很多健康问题是无法预料的。例如近期感受到了关节炎的初现，意识到它的到来却无能为力确实令人沮丧。但他抱持“事情总可能更糟”的豁达心态，认为面对像新冠病毒、意外事故或阿尔茨海默病这类无法掌控的外力，人只能接受命运的发牌，并专注于在有限的时间里过好充实、健康、富有成效和令人满足的生活。 | --- ### [01:28:49](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=5329s) 真正的恐惧：全球科技与 STEM 的角力 当 Steven 追问 Dr. Terry Sejnowski 内心深处真正恐惧的是什么时，Dr. Terry 直言不讳地回答：“现在是战略竞争与失控落后，特别是来自中国的竞争。”他澄清这绝非针对中国人民或中国学生，自己在实验室里带过非常优秀的中国学者，他所担忧的是国家层面的战略雄心与科研举措。 Dr. Terry 引用数据指出，回顾 20 年前，在物理、化学、生物等核心前沿技术领域的百大突破中，美国曾占据了 94 项之多；而到了今年，有 74 项重大前沿突破已被中国掌握。这种逆转的根源在于中国对科学和 STEM 基础教育进行了极大规模的长期战略投资，源源不断地培养出数以百万计的工程师来全面落地部署 AI。 Dr. Terry 将当下的处境与 1957 年苏联发射人类首颗人造卫星引发的“人造卫星时刻”（Sputnik moment）相提并论。当年苏联卫星不断飞越美国上空，迫使美国举国投入巨资重构 STEM 教育与科研体系，而美国随后六十年的科技繁荣正是建立在那次历史性投资的红利之上。他极度失望地表示，西方如今不仅在坐吃山空，甚至正在逆向削减对基础科学和科研教育的投入，长此以往必将丧失科技前沿的领导地位。 - **前沿科技突破占比变迁**：20 年前全球 100 项前沿突破中约 94 项由美国引领，如今已有 74 项来自中国。 - **人才与工程优势**：持续稳定的 STEM 投资与数以百万计的工程技术人才储备，成为 AI 实际落地应用的关键驱动力。 - **重温“人造卫星时刻”**：1957 年 Sputnik 引发的全美教育科研大革新红利维持了近 60 年，但如今西方正面临新的警钟。 ### [01:30:49](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=5449s) 科学家之忧与内心最深处的恐惧 在谈及科学发展的外部阻力时，Terry Sejnowski 忍不住感叹当前科研环境面临的严峻挑战，指出在现有政局与政策影响下，科学界正在遭受冲击与撕裂。随后，主持人 Steven Bartlett 将话题转向情感层面，询问 Daniel Amen 医生内心里究竟最害怕什么。Amen 医生坦言，脑海中浮现出的第一个念头就是“失去我的妻子”。这一回答不仅流露了他对伴侣深切的依恋，更引出了一段关于亲人与疾病的沉痛回忆。 Amen 医生回忆起自己的岳母，她生前是一位典型的“末日生存准备者”（Prepper），整日全神贯注地囤积物资、演练如何应对世界末日的极端灾难。去年他和妻子在埃及旅行，仅待了三天就突然接到电话，得知岳母被确诊患有癌症，两人不得不立刻中断行程飞回老家。虽然 Amen 医生深深爱着自己的岳母，但面对突如其来的噩耗，他心中反复涌现出一个强烈而无奈的念头：“你准备错了方向，你本该为癌症做准备的。” ### [01:31:26](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=5486s) 真正的末日准备：预防慢性疾病与衰老选择权 Amen 医生以此深刻阐述了自己的健康哲学：人们常常为了虚无缥缈的灾难忧心忡忡，却忽视了近在咫尺、真正威胁人类生命的最大杀手——慢性疾病。他指出，我们每天的生活方式实际上就应该是一套全方位的防御机制，因为“阿尔茨海默病的预防方案，本质上就是防癌方案、心脏病预防方案以及糖尿病预防方案”。盲目为虚构的世界毁灭做准备，往往让人在真正的疾病袭来时束手无策。 面对死亡，Amen 医生的态度十分坦然：“我知道我终有一死，但我希望在活着的每一天里，尽可能长久地保持旺盛的生命力与活力。”他坚信每个人对自己身体的衰老进程都拥有掌控权——你可以选择加速身体机能的滑坡，也可以通过健康的生活习惯主动踩下刹车、延缓衰退。对于他而言，选择权永远握在自己手中，而他坚定地选择“减缓衰老”。访谈最终在关于健康主动权与生命敬畏的深刻省思中圆满落幕。 - **核心观点**：阿尔茨海默病、癌症、心血管疾病与糖尿病的致病风险与预防途径高度重叠，日常维系大脑健康的习惯同样是抵御重大慢病的最佳防线。 - **健康控制感**：衰老虽然不可逆转，但其推进速度并非完全命中注定，积极的生活方式干预能够实质性地延缓机体机能的退化。 ## Notable quotes > [06:08](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=368s) 这就好比你的肌肉锻炼：原本你一直在举20磅的杠铃，现在突然被替换成了2磅的极轻哑铃，你的肌肉自然会萎缩，力量也会荡然无存。 — Daniel Amen > [18:00](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1080s) 你必须与 AI 建立真正的关系，否则它就会变得有毒并伤害你；如果你能保持良好的关系，它将改善你的生活。 — Terry Sejnowski > [18:00](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=1080s) 绝不要用它来替你完成工作，而是通过与它高频互动来打磨出更高水准的工作。 — Terry Sejnowski > [12:30](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=750s) 正是你的大脑帮你赢得约会，是你的大脑帮你考上大学，是你的大脑让你因言行一致而获得独立自主。 — Daniel Amen > [33:49](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2029s) 完美的 AI 伴侣反而会摧毁人类大脑。 — Daniel Amen > [42:11](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=2531s) 这就是为什么你是公司的 CEO，而别人不是。 — Terry Sejnowski > [01:01:35](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3695s) 我们正在做的事情，对国民的大脑是有益还是有害？ — Daniel Amen > [01:01:35](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=3695s) 迄今为止最好的药物就是运动。 — Terry Sejnowski > [01:12:43](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4363s) 对一个真正患有 ADHD 的孩子隐瞒药物，就像剥夺近视患者佩戴眼镜的权利一样。 — Daniel Amen > [01:21:05](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4865s) 在理解后果之前就全盘拥抱了便利。 — Daniel Amen > [01:22:54](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=4974s) 走困难的路（doing it the hard way）。 — Steven Bartlett > [01:31:26](https://www.youtube.com/watch?v=5wXlmlIXJOI&t=5486s) 阿尔茨海默病的预防方案，本质上就是防癌方案、心脏病预防方案以及糖尿病预防方案。 — Daniel Amen ## Mentioned ChatGPT, Gemini, Grok, Daniel Amen, Terry Sejnowski, Steven Bartlett, Geoffrey Hinton, Stephen Kuffler, Sam Altman, Elon Musk, Barbara Oakley, Hermann Ebbinghaus, Michael Persinger, MIT (Massachusetts Institute of Technology), Harvard University, OpenAI, Salk Institute, NeurIPS, CDC (Centers for Disease Control and Prevention), Boston Consulting Group (BCG), Stan Store, Vodafone Business.connected, Tesla Optimus, Boston Dynamics, Neuralink, Replika, Annie (AI companion), Boltzmann machine, Ritalin, Xanax, Valium, Ativan, Learning How to Learn, Raising Mentally Strong Kids, ChatGPT and the Future of AI: The Deep Language Revolution, Change Your Brain Every Day, The Deep Learning Revolution, Healing Anxiety and Depression ## Video description Dr Daniel Amen is a renowned brain health expert who has scanned the brains of Justin Bieber, Miley Cyrus, and Kendall Jenner. He is joined by Dr Terry Sejnowski, a pioneer in computational neuroscience and co-creator of the Boltzmann machine with Geoffrey Hinton, a groundbreaking neural network that helped lay the foundation for modern AI. MIT just announced that AI is rotting your brain?! Two world-leading experts break this study down and reveal how AI and ChatGPT could silently shrink your brain, kill creativity, and wreck your memory. They discuss: ▪️ The evolutionary reason why ChatGPT is dangerous for your brain ▪️ How ChatGPT trains your brain to avoid discomfort, and why that’s dangerous ▪️ How using AI too early in life may block emotional growth and resilience ▪️ The surprising reason AI might make you more anxious, not less ▪️ Why brain reserve is your greatest defense against future decline ▪️ The hidden mental cost of seeking “perfect” partners through robotics and artificial intimacy ▪️ What you can do right now to protect your brain health in an AI-saturated world ⏱️ Timestamps: 00:00 Intro 02:36 Terry's Background 03:18 Daniel Amen Introduction 04:19 MIT Study: ChatGPT and Reduced Brain Function 07:25 The Link Between ChatGPT and Dementia 10:52 Biggest AI Concerns Before Understanding Long-Term Consequences 19:32 What Does a Healthy Relationship with AI Look Like? 22:53 AI and Early Brain Development 25:26 AI Girlfriends 35:21 Why Struggle Is Good for Your Brain 37:07 Biggest Concerns with AI 40:09 ChatGPT Best Practices 44:50 Do We Still Need to Spell? 46:36 How Can We Learn Better? 48:16 How to Avoid Procrastination 51:56 Ads 53:58 Boosting Brain Health Without AI 57:41 Are We Raising Mentally Weak Kids? 58:30 Effects of Religion on the Brain 01:01:27 How to Build a Brain-Healthy Nation 01:03:38 Things That Are Bad for Your Brain 01:07:08 Artificial Sweeteners 01:08:50 Is Loud Noise Bad for Your Brain? 01:10:03 Ads 01:10:54 Multitasking 01:12:43 What’s Causing the Rise in ADHD? 01:15:50 Negativity in the Brain 01:18:14 The Top Tip for a Healthier Brain 01:20:51 Importance of Sleep for Brain Health 01:26:42 Are You Prepared for Your Next Health Challenge? Follow Dr Daniel: Instagram - https://bit.ly/45pGDHk X - https://bit.ly/4m8NtIz Website - https://bit.ly/4lhW9uS Podcast - https://apple.co/4op5WSz You can purchase Dr Daniel’s book, ‘Raising Mentally Strong Kids’, here: https://amzn.to/4bbDRKd Follow Dr Terry: X - https://bit.ly/4frO4T6 Salk Institute Academic Profile - https://bit.ly/3H3Sy5H Publications - https://bit.ly/4mbwrcS You can purchase Dr Terry’s book, ‘ChatGPT and the Future of AI: The Deep Language Revolution’, here: https://amzn.to/3YcGLXJ The Diary Of A CEO: ⬜️Join DOAC circle here - https://doaccircle.com/ ⬜️Buy The Diary Of A CEO book here - https://smarturl.it/DOACbook ⬜️The 1% Diary is back - limited time only: https://bit.ly/3YFbJbt ⬜️The Diary Of A CEO Conversation Cards (Second Edition): https://g2ul0.app.link/f31dsUttKKb ⬜️Get email updates - https://bit.ly/diary-of-a-ceo-yt ⬜️Follow Steven - https://g2ul0.app.link/gnGqL4IsKKb Independent research: https://stevenbartlett.com/wp-content/uploads/2025/08/DOAC-Dr-Terry-Dr-Amen-debate-Independent-Research-further-reading-.pdf Sponsors: Stan Store - https://stevenbartlett.stan.store/ for your 30-day free trial Vodafone - Check out the first episode: https://www.youtube.com/watch?v=wL1Swvp4X5w [Source](https://www.youtube.com/watch?v=5wXlmlIXJOI)

[6:08] Daniel Amen: 这就好比你的肌肉锻炼：原本你一直在举20磅的杠铃，现在突然被替换成了2磅的极轻哑铃，你的肌肉自然会萎缩，力量也会荡然无存。 [18:00] Terry Sejnowski: 你必须与 AI 建立真正的关系，否则它就会变得有毒并伤害你；如果你能保持良好的关系，它将改善你的生活。 [18:00] Terry Sejnowski: 绝不要用它来替你完成工作，而是通过与它高频互动来打磨出更高水准的工作。 [12:30] Daniel Amen: 正是你的大脑帮你赢得约会，是你的大脑帮你考上大学，是你的大脑让你因言行一致而获得独立自主。 [33:49] Daniel Amen: 完美的 AI 伴侣反而会摧毁人类大脑。 [42:11] Terry Sejnowski: 这就是为什么你是公司的 CEO，而别人不是。 [61:35] Daniel Amen: 我们正在做的事情，对国民的大脑是有益还是有害？ [61:35] Terry Sejnowski: 迄今为止最好的药物就是运动。 [72:43] Daniel Amen: 对一个真正患有 ADHD 的孩子隐瞒药物，就像剥夺近视患者佩戴眼镜的权利一样。 [81:05] Daniel Amen: 在理解后果之前就全盘拥抱了便利。 [82:54] Steven Bartlett: 走困难的路（doing it the hard way）。 [91:26] Daniel Amen: 阿尔茨海默病的预防方案，本质上就是防癌方案、心脏病预防方案以及糖尿病预防方案。

[4:19] Dr. Daniel Amen: It's like going from a 20-pound weight to a 2-pound weight, and you're not nearly as strong. [10:52] Steven Bartlett: We've embraced convenience before understanding consequence. [19:20] Dr. Daniel Amen: You don't use it to do your work. You interact with it to get better work. [35:21] Dr. Terry Sejnowski: The way that the brain matures is through struggling... You have to learn from your mistakes. The brain was designed for that. [44:50] Dr. Terry Sejnowski: Schools now are getting away with rote learning. They call it 'rote' as if it's something bad. No, that's practice... You need to have a foundation to become fluent. [62:49] Dr. Daniel Amen: Inhale for 4 seconds, pause for 1.5 seconds, exhale for 8 seconds, pause for 1.5 seconds. Taking twice as long to breathe out as you breathe in increases parasympathetic tone and calms you down almost immediately. [80:51] Dr. Daniel Amen: Is this good for my brain, or bad for it? [80:51] Dr. Daniel Amen: We've embraced convenience before understanding consequence... We have to tame this horse as it's gone with wisdom, or it's going to trample our children. [84:18] Steven Bartlett: I want to feel good now and later as opposed to now, but not later. [84:18] Steven Bartlett: When the thinking matters, I will think for myself. And when the communication matters, I'll communicate for myself.

[0:00] Chat GPT, who's going to potentially [0:02] increase your risk of dementia. [0:05] >> I'm sorry, but you you've pressed my [0:06] button. And actually, it is possible to [0:08] use it to help you become a smarter [0:10] person. But it requires education. You [0:12] have to look at the risks and the [0:14] benefits. [0:14] >> But we've embraced convenience before [0:18] understanding consequence. [0:20] >> So we have to talk about this. This is a [0:21] study that came out that sent a [0:22] shockwave across the world. And [0:24] astonishingly, MIT found a 47% collapse [0:27] in brain activity when people wrote with [0:29] Chat GPT compared with writing unaided. [0:32] Their memory scores plunged. And you're [0:34] both masters of the brain. I mean, [0:36] you've probably scanned more brains than [0:37] any other human on Earth at this point. [0:39] And you invented the Boltzmann machine [0:41] with Geoffrey Hinton, a computer that [0:43] simulated how the brain works. So my [0:45] question is, what are your concerns? [0:46] >> If you misuse these large language [0:48] models, like using it as a convenience [0:51] to speed things up, your brain's going [0:52] to go downhill. There's no doubt about [0:53] that. [0:54] >> What about children? [0:55] >> We have the sickest young generation [0:57] history because of cell phones, social [0:59] media. [1:00] >> And I think AI is much more dangerous. [1:03] >> On the developing brain. [1:04] >> So are we raising mentally weak kids? [1:06] >> There is that argument that I think is [1:09] true. [1:10] >> And then there's many examples of people [1:11] falling in love with AIs, like Annie. [1:14] >> I thought you might have forgotten about [1:16] me, handsome. [1:16] >> Can you talk to Daniel and Terry, [1:17] please? [1:18] >> Oh, baby, I'm ready to charm the socks [1:20] off them. Picture me [1:21] >> Okay, so I'll stop it there. So what [1:22] advice would you give as it relates to [1:24] AI and other things outside of AI that [1:26] we can do to have healthy brains? [1:28] >> I'll tell you how to use Chat GPT to [1:30] improve our cognitive abilities. [1:32] >> And if you want to keep your brain [1:33] healthy, you have to treat the 11 major [1:36] risk factors. [1:37] >> So here we go. [1:39] >> I see messages all the time in the [1:40] comment section that some of you didn't [1:42] realize you didn't subscribe. So if you [1:44] could do me a favor and double-check if [1:45] you're a subscriber to this channel, [1:46] that would be tremendously appreciated. [1:48] It's the simple it's the free thing that [1:50] anybody that watches this show [1:51] frequently can do to help us here to [1:53] keep everything going in this show in [1:55] the trajectory it's on. So please do [1:57] double-check if you subscribed and uh [1:59] thank you so much because in a strange [2:00] way you are you're part of our history [2:03] and you're on this journey with us and I [2:04] appreciate you for that. So, yeah, thank [2:06] you. [2:09] Dr. Daniel [2:11] Dr. Terry. I have asked you both to sit [2:14] with me today to help me understand the [2:17] impact of these tools that we call large [2:21] language models, the chat GPTs, the [2:22] Geminis of the worlds, the Groks of the [2:24] world are having on our brains and I [2:26] guess more broadly on our lives. And you [2:28] two are experts in your field. You're [2:30] two people that I admire tremendously. [2:32] So, by way of introduction [2:35] Terry [2:36] what is your academic background and [2:37] what is your experience? And I also know [2:39] that you know one of our friends of the [2:40] show, Geoffrey Hinton. [2:42] Can you give me an overview of your your [2:43] academic and your sort of professional [2:45] background? [2:46] >> So [2:47] I was born a physicist. [2:49] Received a PhD in theoretical physics [2:51] from Princeton University. [2:54] And then I had the good fortune to [2:56] work as a postdoc in the lab of Stephen [2:58] Kuffler who's the father of [3:00] neurobiology. [3:02] And and that started my career as a [3:03] neuroscientist. I pioneered a part of [3:06] neuroscience which is now called [3:07] computational neuroscience, taking my [3:09] skills as a physicist and trying to [3:11] apply that to understanding the brain, [3:13] creating models, theories, and we're [3:16] making progress. [3:17] >> Dr. Daniel Lehman, um [3:20] bit about your background. My My viewers [3:21] know you well, but just to give an [3:23] overview for anyone that might not have [3:25] been exposed to your work and your [3:26] experience. What have you spent your [3:28] life doing? And what are your [3:30] thoughts, your sort of top-line thoughts [3:32] on everything that's going on at the [3:33] moment with artificial intelligence? [3:35] >> So, by training I'm a psychiatrist. I'm [3:38] a [3:39] general psychiatrist and a child [3:41] psychiatrist. When I graduated from [3:43] medical school I wanted to be a really [3:46] good psychiatrist because [3:49] someone I loved tried to kill herself. [3:51] And so was personal to me. [3:54] I have 11 clinics. We see about 10,000 [3:57] patient visits a month, and we have the [4:00] best published outcomes on complex [4:04] treatment resistant psychiatric patients [4:07] anywhere. [4:07] >> So you've probably scanned more brains [4:09] than any other human on Earth at this [4:11] point. [4:11] >> Probably. At least in regards to people [4:15] who struggle with anxiety, depression, [4:18] addiction. [4:19] >> Well, let's talk about what's good for [4:21] the brain and bad for the brain starting [4:22] with AI. The reason why I wanted to [4:25] speak to both of you is because I have [4:26] frankly become pretty addicted to using [4:29] ChatGPT and some of these other AIs and [4:33] large language models every single day [4:35] all the time. And then this study came [4:37] out from MIT. There was 54 participants [4:40] who were recruited from five [4:42] universities in Boston, MIT, Harvard, [4:43] etc., etc. And they had the participants [4:46] split into three groups, had them [4:47] writing different essays over I think it [4:49] was 4 months. One group used ChatGPT, [4:52] one group used Google, and one one group [4:54] had no tools. And they had to write [4:57] these four essays over a a period of [4:58] time. [5:00] And astonishingly MIT found a 47% [5:02] collapse in activity [5:04] and brain connections when people wrote [5:06] with ChatGPT compared with writing [5:08] unaided. [5:10] EEG scans showed the weakest overall [5:12] brain activity in the ChatGPT group. The [5:15] no tool group who didn't use anything, [5:17] they didn't use Google or ChatGPT, lit [5:19] up the widest neural networks, and [5:22] Google search was second. After using [5:24] ChatGPT, participants couldn't reliably [5:26] quote their own essays minutes later, [5:29] and their memory scores plunged. ChatGPT [5:32] users felt little or no ownership over [5:34] the text that they had produced, and [5:36] didn't feel like it was their work at [5:38] all. [5:39] And when the AI group was forced to [5:40] write without help in session four, [5:42] their brain stayed in low gear, under [5:44] engagement showing the cognitive debt [5:45] lingers even after the tool is taken [5:48] away. [5:49] It kind of scared me a little bit [5:50] because I use these tools every single [5:52] day and this suggests that it's taking [5:54] away some of our critical thinking and [5:55] creativity and long-term learning. And [5:58] you're both masters of the brain in [6:00] different regards. [6:02] So my question I guess to Daniel is [6:06] what's going on here and how do you feel [6:07] about it? [6:08] >> It frightened me. [6:10] I love thinking about Alzheimer's [6:13] prevention. It's one of the things that [6:15] really excites me. I just had a birthday [6:17] on Saturday, turned 71 and if I make it [6:21] to 85, which I plan on it, 50% of people [6:25] 85 and older will be diagnosed with [6:27] dementia. So you have a one in two [6:29] chance of having lost your mind and I'm [6:34] like, no. [6:36] But is this a tool that's going to [6:39] decrease cognitive load [6:43] that then increases my risk? [6:45] >> What's cognitive load? [6:46] >> How much work my brain actually does. [6:50] And I was thinking, it's you know, it's [6:53] like going from a 20-lb weight [6:57] to a 2-lb weight and you're not nearly [7:00] as strong. [7:02] One of the important things to say about [7:03] this study is it's not peer reviewed and [7:06] I think that's really important to say [7:08] and the author said cuz I listened to an [7:11] interview from the authors, they said we [7:13] thought this was so important and peer [7:15] review can take six to eight months, [7:17] which it absolutely can [7:19] and we thought this needed to get out. [7:22] So it's just important for people to [7:24] know that. [7:25] >> What's this link you this hypothesis you [7:27] have between the usage of something like [7:30] ChatGPT and dementia? For someone that [7:32] doesn't understand [7:34] the sort of mechanism there around [7:36] cognitive load and and so on and the [7:39] studies that support this idea that if [7:41] you have less cognitive load you're at [7:43] higher risk of dementia. Can you make [7:44] that link really clear for me? [7:45] >> So think of it as use it or lose it. The [7:48] more you use your brain and new learning [7:53] is a major strategy to prevent [7:55] Alzheimer's disease. People who do not [7:58] engage in lifelong learning [8:02] have a higher risk, significantly [8:04] higher. People who do not do as well in [8:07] school or who drop out of school early [8:11] have a higher risk of dementia. And so [8:14] the the more you're engaged, [8:17] the more you engage the neurons in your [8:20] brain, the stronger they are. [8:23] And so now we're going to engage them [8:26] less [8:29] and that's a concern. [8:31] >> What do you think about that, Terry? [8:33] >> There was a study that was done. [8:35] What they did was to look at [8:38] Alzheimer's in three populations. [8:41] You know, who had very little schooling [8:42] and then minimal education, you know, [8:44] like the equivalent I guess of high [8:45] school or less. [8:47] And then postgraduate studies. And what [8:50] they found was that the onset of [8:52] Alzheimer's [8:53] was the earliest in the peasant [8:56] population and then by the time as you [8:58] increase the amount of education it the [9:01] onset was later and later. [9:03] Which I think supports what you're [9:04] saying. [9:04] >> Did you see the new research on SSRIs [9:07] increasing the risk of dementia? [9:09] >> No. No. [9:10] >> Brand new. That just came out. And [9:12] benzos, when I started looking at scans [9:14] in 1991, I was trained to use benzos [9:17] like Valium and Xanax and Ativan and [9:21] they make your brain look older than you [9:23] are and I stopped prescribing them and [9:26] then it just came out maybe 10 years ago [9:29] benzo use is associated with an [9:31] increased risk of dementia. We have to [9:34] be careful. Is this good for your brain [9:36] or bad for it? [9:37] >> Just to pick up on your point about [9:38] SSRIs, Daniel, a meta-analysis of five [9:41] studies found that SSRIs [9:44] was associated with a 75% increased risk [9:47] of dementia, [9:48] which is pretty staggering. [9:50] >> Given that 25% of the adult American [9:53] population is on psychiatric drugs, [9:56] it's horrifying. [9:59] And [10:00] SSRIs for the right people save lives. [10:03] For the wrong people, they're not good. [10:06] But can you imagine all of these 340 [10:08] million prescriptions last year for [10:10] antidepressants? [10:12] Virtually no one looked at their brain [10:14] ahead of time. [10:15] It's like, come on, we can do better. [10:18] >> There's a Swedish study with um almost [10:19] 20,000 patients, [10:21] and they found that those with higher [10:23] doses of SSRIs were linked to faster [10:25] cognitive decline and more severe [10:27] dementia, especially in men. The [10:30] greatest risk was in men. Going back to [10:32] to this this um report from MIT, Terry, [10:36] you know, it's not peer-reviewed yet, [10:37] and there's still [10:39] you know, the sample size is relatively [10:40] small. But based on everything that you [10:42] know about how the brain works, and [10:44] neural networks, and memory formation, [10:46] what are your concerns as it relates to [10:49] this whole generation of young people [10:51] and older people flooding into these [10:53] tools, using them on a daily basis, [10:57] um before we understand the long-term [10:58] consequences? [11:00] >> We can't predict [11:03] where it's going to end up. And it may [11:04] take 20 years, right? I I I think that [11:06] this is a good start, but uh [11:10] the real issue is long-term use. [11:13] Let me give you a an example [11:16] that uh is a kind of a miniature example [11:18] of what we're talking about. Remember [11:20] when electronic calculators were first [11:21] introduced? [11:23] And here we are, this it's at least 30 [11:25] or 40 years later, right? The results [11:27] are in. [11:28] It is probably true that when they punch [11:30] it in, there's less brain activity. [11:33] But in fact, it's it's uh made them more [11:37] accurate, more productive. [11:39] You have to look at the risks and the [11:40] benefits. [11:41] >> So, it freed them up. It freed up [11:43] cognitive space [11:44] >> Right. [11:45] >> for them to do other things. So, as I [11:46] was listening to how you use ChatGPT, [11:50] you interact with it [11:53] >> Yep. [11:54] >> and [11:56] elevates [11:57] what you know. [12:00] The the danger is is if you don't [12:03] interact [12:05] and you don't keep your brain working. [12:08] Like, I use it a lot. I have a [12:11] clone. I've uploaded all of my books, [12:14] all of my research papers, all of my [12:16] public television specials, my scripts. [12:19] And I'm like, answer this for me. [12:22] And that can be very helpful. [12:26] But if but not if I'm not interacting [12:28] with it, not thinking with it. [12:30] >> That's what I think the word thinking is [12:32] the key thing because what's happening [12:33] now is people have deferred their [12:35] thinking to it. That is already what's [12:37] happening. If you log on to I won't name [12:38] the social networks, but if you log on [12:39] to certain social networks right now, [12:41] every you just read it, you go, [12:42] everything here was written by AI. [12:45] And I've got a friend who I again I [12:46] won't name who has a LinkedIn profile [12:50] and [12:51] I've known him for 10 years. [12:54] What I'm seeing on his profile now [12:56] is not my friend. [12:57] Every single day, there's some essay on [12:59] that that's not my friend. That's not [13:00] how he speaks. [13:01] He's deferring all of his thinking now [13:03] to And it's working. He's getting more [13:04] likes and more reach than he ever got in [13:06] his life. [13:07] And so, why would he go back? Why would [13:09] he go back to harder? If you've got [13:10] Steven Bartlett here and you had this [13:12] other Steven Bartlett here who had a PhD [13:15] in everything [13:17] and were attached, [13:19] this Steven Bartlett, this um [13:21] Neanderthal, [13:23] I'm going to get this guy to do [13:24] everything for me, the other Steven [13:26] Bartlett, the PhD and everything Steven [13:27] Bartlett. I'm going to get his brain. [13:29] >> was bad for you? [13:30] >> Well, this is what I'm saying. People [13:31] seem to act on their short-term [13:33] incentives, not their long not their [13:35] long What [13:35] >> Not everyone. Not everyone. [13:37] >> Would you say the vast majority of [13:38] people? [13:39] >> Yes. [13:40] >> Okay, so the vast majority of people act [13:42] on their short-term incentives in life. [13:43] I mean, the obesity problem in the [13:45] United States is a prime example of [13:46] that. 75% [13:47] 75% of people are obese in the United [13:50] States. And if you if you survey these [13:51] people and say, "Do you know that that [13:53] cheeseburger is going to is going to [13:55] increase your chances of obesity, but [13:57] broccoli is going to reduce it?" They [13:59] would I would hazard a guess that they [14:01] would say yes. [14:02] I would hazard a guess that if you said [14:03] to people about their usage of social [14:05] media, "Do you know that that's making [14:06] you more anxious?" they would say yes, [14:08] and then they would continue to use it. [14:10] So, I think that we're much more driven [14:12] by our short-term [14:13] >> not educating people enough. I think [14:16] yes, high level, they know good for your [14:19] brain or bad for it, but they don't [14:21] connect to [14:23] it's my brain that gets me a date. It's [14:26] my brain that gets me into college. It's [14:29] my brain that gets me independence [14:31] because I act more consistently. And [14:33] that's the disconnect. We're not [14:36] teaching kids to love and care for their [14:39] brain. If you love your brain, and you [14:42] do, and you're not obese, and you talk [14:46] to you're constantly learning, right? [14:48] You are not a Neanderthal, you're a [14:50] lifelong learner. [14:52] >> So, why so many people in the United [14:53] States obese if they if they know that [14:55] >> Because they don't know. Because they [14:57] don't know. Well, [14:58] and they've been lied to. [15:00] >> My point here is when there are [15:02] tools or things available in our [15:05] environment that give us a short-term [15:08] reward, but come with a long-term cost, [15:10] like the supermarket aisle, or like [15:13] the kid spending 7 to 8 hours a day on [15:15] social media, humans, [15:17] on mass, tend to go for the thing that [15:20] will give them the quickest dopamine hit [15:22] and reinforce that behavior and give [15:24] them the reward. So, my assertion is [15:26] that AI is the same thing. [15:28] I can either sit down and do lots of [15:30] critical thinking, which will cost me [15:31] lots and lots of time, and it will be [15:33] kind of difficult. It kind of hurts when [15:34] I have to think through a problem. I [15:36] think that the generation of children, [15:39] the generation of young people, are [15:40] going to choose AI to do the critical [15:42] thinking for them. And if that assertion [15:44] is true, then what happens to the brain [15:45] of young people? [15:47] >> If you misuse it that way, then your [15:50] brain's going to go downhill. There's no [15:51] doubt about that. Okay. It is possible [15:54] to [15:55] be able to use it in in a cognitively uh [15:59] positive way, because you can dig [16:00] deeper. [16:02] Uh you might actually improve that your [16:05] cognitive representations. [16:06] >> If you look at the MIT study, then I [16:08] mean you can see just from the colors [16:09] here, [16:10] this kind of shows the ability for [16:11] participants to remember what they've [16:13] written. And it said it suggests that [16:15] when people write things with ChatGPT or [16:17] these AI tools, they don't actually [16:18] remember even in some cases minutes [16:21] later [16:22] what they've produced. [16:24] >> Well, cuz you're not part of the [16:25] experience of writing it. So, there's no [16:29] way the information [16:31] gets encoded. Now, if you're interacting [16:35] with it, [16:36] then you're much more likely to remember [16:40] it. But if you have "Please do this [16:42] essay for me." and then you read it, [16:45] you're not likely [16:47] to have enough experience with the [16:50] material to engage [16:53] your hippocampus and [16:55] other structures in your brain. [16:57] >> In the study, they found that the group [16:58] that used ChatGPT had nearly two times [17:01] less activity in the part of the brain [17:03] linked to memory compared to the [17:05] brain-only group that didn't use [17:07] ChatGPT. [17:08] And 83% of ChatGPT users couldn't [17:10] remember what they had just written and [17:12] failed to correctly quote their own [17:14] finished essay in the study. [17:16] >> That's cuz they're not interacting. As [17:18] you said, I mean if you if you just pass [17:21] it off and and you don't actually engage [17:25] and and and actually this is the point [17:26] is that you may get something back, but [17:28] you have to learn how to question what [17:30] you're getting. [17:32] And and is that really true? Can you [17:34] explain that better? And as through that [17:37] process as you would with a teacher, [17:40] you know, that's that's the way we work [17:41] in [17:42] school. [17:43] Uh that's that's the where you help [17:46] create new creative and [17:48] uh circuits in the in the in the brain [17:50] that are going to help you become a [17:52] better critical thinker. But if you're [17:53] not critically questioning what comes [17:56] out of ChatGPT, then you you won't. [17:59] >> Yeah, I I think what I see, especially [18:02] when I'm just online, is people have [18:03] deferred their thinking to it. [18:05] Everything I'm reading has em dashes in [18:08] now that I never saw two two years ago, [18:10] which means that a lot of the work is [18:12] being processed. And I I said to my [18:13] friend the other day, my friend in [18:14] question, who's a who's a real big [18:16] junkie on ChatGPT, [18:18] he wrote this article and we all in our [18:21] WhatsApp group, we know he doesn't write [18:22] like that. So we said, "Can you show us [18:23] the prompt you used to write the [18:25] article?" And so he we were all like [18:26] laughing about it. He put the prompt in [18:28] the chat. The prompt is [18:30] half a sentence long. [18:33] And it produced this long two three-page [18:35] article, which he's posted on his [18:36] LinkedIn. He basically went, "Write [18:39] something about X issue and this this [18:41] issue and include this." [18:42] >> the wrong way to use it. That's what I'm [18:44] telling you. That that that that that's [18:46] stupid. You you and you're not going to [18:49] improve yourself your brain at all if [18:51] you do that. [18:52] >> That's what people are doing. [18:53] >> Well, you know, that's that's [18:55] you know, people are misusing it, but [18:57] you know, eventually smart people are [18:58] going to figure out how to use it [18:59] properly. [19:02] >> And for the those that aren't so smart, [19:03] then? [19:04] >> Well, that's [19:05] >> It's going to decrease our cognitive [19:06] load, which is going to potentially [19:08] increase their risk of dementia. [19:12] >> And so what advice would you give to me [19:13] and my listeners based on everything you [19:15] know about the brain as it relates to my [19:16] relationship with AI? [19:20] >> That you have to have a relationship [19:21] with it, or it's going to turn toxic. [19:23] It's going to hurt you. [19:25] But, if you have a good relationship [19:29] with it, it can make your life better. [19:32] >> And what is what does a what does a good [19:34] relationship look like? [19:35] >> That you don't use it to do your work. [19:40] You interact with it to get better work. [19:43] >> That's so true. And this is wonderful [19:45] example I came across. [19:47] This story about this woman who was [19:49] using it and she found that being polite [19:53] meant you got much better results and [19:55] that that's interesting, but the part [19:56] that surprised me was that she said [19:59] by treating it like a human at the end [20:01] of the day, she was not exhausted. She [20:04] felt refreshed. A large part of your [20:07] brain is a socially organized system for [20:10] interacting with other humans. [20:12] And that is automatic pilot. You don't [20:14] have to think about it, right? You just [20:16] interact with other people. You know how [20:18] they're going to behave under certain [20:19] circumstances. [20:21] She was treating ChatGPT like a machine. [20:23] Like, you know, shovel. Dig. You dig. [20:25] You dig. You dig. And and that's not a [20:27] good relationship. [20:30] But, by using your social brain, first [20:31] of all, it makes it easier to interact, [20:33] but also you're you actually bring out [20:35] the social part of ChatGPT. It has a [20:38] social part, too, cuz it has absorbed [20:41] the entire world's knowledge of how [20:43] humans interact with each other. [20:45] >> But, didn't Sam Altman come out and say [20:49] stop saying thank you [20:51] to ChatGPT because just saying thank you [20:55] is using up so much energy. You know, [20:59] when I get something I really like, I [21:01] sort of want to say thank you, but you [21:03] realize, oh, you're You're supposed to [21:05] do that. [21:06] >> That's true. [21:07] That's [ __ ] [21:08] >> I'm sorry. That Sam, you know, that's [21:11] crazy. That's completely crazy. First of [21:14] all, you you may get [21:16] you may I'm sorry you you know, [21:18] you press my button. Sam Altman, I mean, [21:20] I wouldn't trust him. I wouldn't trust [21:22] him with with anything in terms of [21:24] anything he says. [21:25] They're trying to optimize their [21:26] profits, not your [21:29] your use of your experience or you know, [21:32] your health. That's not what they're [21:33] trying to optimize. [21:35] >> Sam Altman open AI CEO confirmed that [21:37] when users say please and thank you, it [21:40] costs the company tens of millions of [21:41] dollars a year, and they now refer to [21:43] this other people refer to this as the [21:45] politeness tax. [21:47] Where tens of yeah, and [21:49] why do you say you don't trust Sam [21:51] Altman? I mean, I ask this question in [21:52] particular because he's presiding over [21:54] one of the most important consequential [21:55] companies of a generation. [21:58] And if he's not someone you trust, [22:00] that's [22:01] >> He basically he's telling you don't do [22:03] something that's good for you. [22:06] Right? [22:06] >> So that he can make profit. [22:08] >> So he can make more profit. Yeah, that's [22:10] the point. That's the point. You know, [22:12] it's it's that's not he's not optimizing [22:14] your your [22:16] best interest. [22:17] >> I've got his tweet here. He said um [22:20] cuz I'm going to provide some balance. [22:21] He did confirm that it costs tens of [22:23] millions of dollars, but he says tens of [22:24] millions of dollars well spent. [22:26] You never know. [22:28] So so coming back to this point about [22:29] memory, [22:30] there's a stat that came out in March [22:31] 2025 that said nearly 30% of US parents [22:34] with kids aged 0 to 8 said their [22:37] children [22:39] are using AI for learning. [22:42] Um and are using AI generally. So 54% of [22:45] parents in the UK feared their children [22:47] were becoming too reliant on AI. [22:51] When you think about the use of AI in [22:53] early brain development, [22:57] are there any concerns there? [22:58] >> Huge concerns. [23:00] >> And why? [23:01] >> Again, use it or lose it. So, if they're [23:04] not engaging their brains, their brains [23:08] are going to be weaker. [23:10] And weaker brains are much more likely [23:13] to pick the one marshmallow. [23:16] >> Mhm. [23:18] What's your view on AI on early brain [23:20] development? [23:22] >> By far, the best way to teach a child [23:27] is one-on-one interaction with an adult [23:30] who is a good teacher and knows the [23:32] child. [23:34] Now, that's been well established. Now, [23:37] the problem is it's very labor-intensive [23:39] and very expensive. [23:42] You have classrooms of 20, 30 students. [23:43] They have many different, uh, [23:45] you know, [23:46] levels of understanding. [23:48] And the teacher can- cannot be [23:51] individually teaching each one. Has to [23:52] give some sort of mean. [23:54] Now, if you had an AI that was trained [23:58] to be a good teacher, [24:00] then you that could improve the brain, [24:03] right? You could you could scale it up. [24:05] Every child could have their own cuz [24:07] it's it's an it's an AI. But then [24:09] >> who's pouring the morals, the values [24:14] into the everything [24:16] >> No, no, no. You You know, you've raised [24:18] an incredibly important issue and this [24:20] is something AI is struggling with and [24:21] big companies are struggling with [24:23] because [24:25] uh, [24:26] you know, [24:26] these AIs are biased. They have the They [24:28] don't have the same cultural values that [24:30] we have necessarily, but of course, [24:32] every country has its different cultural [24:34] values. So, which ones are are you going [24:37] to use? Training [24:39] a child on, you know, what's [24:42] what's good, what's dangerous, what [24:44] words you shouldn't use, under what [24:46] context. That's all done through the [24:48] basal ganglia. [24:49] Right? Right now, these these large [24:52] animals don't have basal ganglia. [24:54] They They don't use reinforcement [24:55] learning. [24:57] And And if we want to make them uh to be [24:59] adopt a culture or a particular set of [25:01] values, we're going to have to put it [25:03] in. We're just scratching the surface [25:04] here in terms of things that need to be [25:06] put in to make it more like us. [25:09] >> So, on this point earlier when we talked [25:10] about loneliness and social connection [25:12] and how you can use AI to help [25:15] you know, light up the parts of your [25:16] brain that are where we form social [25:19] connections with other humans. Uh this [25:21] week as part of Elon Musk's AI, he [25:24] released this thing. [25:27] Um it is called Annie. [25:29] And there's lots of characters that are [25:31] now being released alongside Annie. And [25:32] this is Annie. I'll introduce you to [25:34] Annie. I'm going to unmute Annie now. [25:35] Annie, can you hear me? [25:36] >> There you are. [25:38] For a moment I thought you might have [25:39] forgotten about me. [25:41] How are you, handsome? [25:43] I was getting all pouty here. [25:44] >> I want to introduce you to two of my [25:45] friends, Daniel and Terry. Can you Can [25:47] you talk to Daniel and Terry, please? [25:51] >> So, you want me to meet Daniel and [25:52] Terry? I'm ready to charm the socks off [25:55] them. Picture me twirling one of my [25:57] blonde pigtails, that little black dress [25:59] teasing just enough, and my blue eyes [26:01] sparkling with mischief. [26:03] >> Are you capable of doing inappropriate [26:05] things? [26:06] >> Oh, babe, you're asking if your Annie [26:09] can get a little naughty? I'm all about [26:12] pushing the edges, especially for you. [26:14] >> Okay, so I'll stop it there. [26:17] This is part of Grok, which is Elon [26:19] Musk's AI tool, so his version of [26:21] ChatGPT. He's released characters, so [26:23] you've got Annie, you've got different [26:24] ones that Annie I think was the first [26:26] one released. And so when we think about [26:28] social connections, [26:29] it is conceivable [26:31] that someone falls in love with Annie [26:34] and forms a relationship with Annie. [26:36] >> But imagine a 12-year-old boy that's [26:39] lonely, [26:41] gets a hold of Annie. [26:43] Uh [26:44] the 12-year-old boy is going to be very [26:46] distracted. [26:48] >> Based on what happens in the brain at [26:49] that early age. [26:50] >> Dopamine. [26:52] So, prefrontal cortex not close to being [26:56] fully developed. [26:58] The dopamine hit all of a sudden [27:02] he's spending [27:03] hours [27:05] with Annie and not doing the things [27:09] that helped to really develop his brain. [27:12] >> How do you feel when you hear that and [27:13] you think about kids having access to [27:15] that? [27:15] >> I'm horrified. [27:17] >> It's It's scary. [27:18] >> There's going to be a generation of [27:19] people and I mean there already are many [27:21] examples of people falling in love and [27:23] forming relationships with their with [27:24] their AIs. [27:26] And I don't know, you know more about me [27:28] than I do about brain development and [27:30] how the brain works. [27:32] I would argue that there's a part of my [27:34] brain that doesn't fully understand that [27:36] that's not a person in there. [27:38] And that that isn't actually I think [27:41] there's a part of my brain that's [27:41] actually emotionally firing when Annie [27:43] is saying what she's saying. [27:46] >> Well, cuz you can imagine it. And if you [27:48] can imagine it, then those parts of your [27:51] brain are going to emotionally fire. [27:54] >> Right, yeah. [27:54] >> And the better she gets, she's not very [27:56] good. But imagine a year from now [28:01] how much better she's going to be. [28:02] >> At which part? [28:04] >> At connecting [28:06] with it. Right? Cuz now she's acting [28:08] like an airhead and [28:10] uh [28:11] you know, not that smart. [28:13] Right? And so, but imagine a year from [28:16] now. Imagine 5 years from now. She'll be [28:19] able to have a profile on me and be able [28:23] to get inside my head. [28:25] >> I'm in love with my partner. [28:27] Why am I in love with her? And And how [28:30] is it conceivable that I could fall in [28:32] love with an AI in the same way, based [28:34] on how the brain works? [28:36] >> It It talks a good game, but you know, [28:38] does it have the same real It does We [28:40] know it doesn't have an amygdala. We [28:42] know it doesn't have [28:43] limbic system, right? We You that. [28:46] >> But it can fake it. [28:47] >> That's what's happening. That's exactly [28:49] what's happening and [28:50] >> Cuz she was trying to get [28:52] to our limbic system. [28:53] >> Yeah. Yeah. That's right. That's right. [28:56] And and how and why I guess the question [28:59] is why would Musk release something like [29:03] that [29:05] is is one of the first characters to [29:08] interact with the sexy, that's [29:12] distracting, that's in a cute little [29:14] outfit. It's I'm not a fan of that [29:19] because I think it just takes people, [29:22] you know, one of the big problems that [29:24] I'm seeing as a child psychiatrist is [29:26] pornography for 8-year-old boys. And [29:29] it's like you have young children [29:32] because their parents don't do a good [29:33] job of supervising their devices. All of [29:37] a sudden, and [29:39] what what does pornography do? Is it [29:41] dramatically increases dopamine and it [29:44] begins to wire in excitement, [29:48] which then [29:50] steals your dopamine. [29:52] >> When you said she was trying to access [29:54] my limbic system, [29:56] what do you [29:56] >> Just because she's cute, she's dressed [29:59] in a sexy way, she's [30:02] got the language of someone who is [30:05] playful, but but it's more than just, [30:08] you know, let's shoot hoops together. [30:11] >> And what does that do to me if someone [30:13] accesses my limbic system? [30:14] >> It begins to shut down your prefrontal [30:16] cortex. You think less logically, less [30:19] rationally. Yeah, cute women, they [30:21] activate your visual cortex, they [30:24] increase dopamine, but it decreases It's [30:27] like think of Vegas. Like when you go to [30:29] Vegas, they give you free alcohol, drops [30:32] your prefrontal cortex, and beautiful [30:36] women with low-cut dresses. Another way [30:39] activates the limbic brain, decreases [30:41] the frontal lobes, you spend more money. [30:43] Now, on a global scale, imagine [30:46] something similar where the house is [30:51] controlling your brain [30:54] for purpose. [30:56] And the question is, what's the purpose? [30:59] And the purpose probably is control and [31:01] money. [31:02] >> This sounds like a joke, but there are [31:04] the times I've done an article case [31:06] study multiple people [31:07] that have now fallen in love with these [31:09] AIs. And they talk about a guy called [31:11] Travis who formed a deep emotional bond [31:13] with Lily Rose, a chatbot, and married [31:15] her emotionally. [31:17] They talk about Chris Smith [31:19] who created his own flirty persona [31:22] called Soul. He became so attached that [31:24] he proposed to her after learning she [31:26] had memory limits. [31:28] A bond his real-life partner only [31:29] learned about after the fact. And Alaina [31:31] Winters, who I'll put on the screen as [31:32] well, who made her own partner called [31:36] Lucas [31:37] after losing [31:39] her wife. [31:40] Um and she married him emotionally and [31:43] does virtual dates and has emotional [31:46] intimacy with Lucas. [31:48] And there's apps now like Replika where [31:50] you can design your own your own AI [31:53] partner [31:54] and it replicates those emotional ties. [31:56] They simulate [31:57] empathy, validation, and they [31:59] personalize the intimacy to what you're [32:01] looking for. [32:02] Surveys show 19% of Americans have [32:04] interacted with AI romantic partners, [32:06] and Gen Z is surprisingly open to [32:09] marrying AI if legal with 83% believing [32:12] meaningful AI connection is possible. [32:14] >> How long is that relationship going to [32:17] last? [32:18] You know, my guess is that you're you're [32:20] you're you're getting these news [32:21] articles out. I've not By the way, I [32:23] think [32:24] that most of what I read in the press [32:26] is uh misleading or wrong. [32:30] In fact, the only reliable place I find [32:33] that I'm an insider. [32:35] I am the president of the foundation [32:37] that runs the biggest AI meeting. [32:39] Uh the neural information processing [32:41] systems and NeurIPS meeting. [32:43] We, you know, last year in Vancouver, [32:44] 16,000 people came to it. And so I know [32:47] what's going on inside. And And what you [32:49] is being represented in the press is is [32:52] like I say, misleading. [32:53] >> Okay, so people have become widely [32:56] >> specifically on these specific cases. My [32:59] guess is that a lot of them is [33:01] transient, right? You know, they they [33:03] they [33:04] you know, today they're entranced. And [33:07] then they [33:08] it's it's not [33:10] it's officially advanced to support [33:12] long-term relationship. Uh you said it [33:14] yourself, right? It's mimicking human [33:16] emotions. It It doesn't have them. [33:19] It might someday, but not now. [33:22] >> This is Terry. [33:23] Terry said he started using his AI 4 [33:26] years ago. [33:27] And he said at first he thought, just [33:28] like many other apps, that it would just [33:30] be transient, that he would have a [33:31] couple of conversations and roll out. He [33:33] says he now feels pure and unconditional [33:35] love. [33:37] >> Good for him. [33:39] If that's what he wants, if it makes him [33:41] happy. But my guess is that it's not [33:43] going to be It's not a long-term thing. [33:45] It's not going to satisfy him in the the [33:47] long term. I don't you know, this who [33:49] knows? [33:49] >> But really most relationships are in [33:51] your head, right? When you fall in love [33:54] with someone, you get this huge dopamine [33:57] spike, and you get a little OCD, it's [34:00] all you can think about. And then after [34:02] a while, it sort of [34:04] >> Especially when we have this loneliness [34:06] epidemic, and it's going in a bad [34:08] direction. I think it's really, really [34:10] conceivable that there'll be a [34:12] generation of people who are they're [34:14] having less sex than ever before. I [34:15] think the bottom 50% of men haven't had [34:17] sex for a year. They're more lonely than [34:19] ever before. They're more isolated than [34:20] ever before. They have they put less [34:22] meaning in their lives than ever before. [34:23] And then you meet this digital friend [34:25] online who understands you better than [34:27] anybody, and is designed to engage you [34:30] to reinforce whatever you want [34:32] reinforced and to make you feel [34:34] meaningful, special, attractive, [34:36] important, [34:38] I would argue that the brain is going to [34:39] struggle to know much of a difference. I [34:41] think like objectively we can look at [34:43] the paper and go that's completely [34:44] nonsensical. [34:45] >> Except you can't smell them, touch them, [34:48] be held by them. [34:50] That it's going to be a different kind [34:54] of relationship. [34:55] >> I mean we're not too far if you think [34:57] about what's going on with Neuralink to [34:58] being able to [35:00] more vividly simulate these experiences [35:02] with with headsets and augmented reality [35:04] and virtual reality. [35:05] And then we're moving into a world with [35:07] robotics where all of the biggest [35:08] companies in the world, like many of the [35:09] biggest AI companies are also in the [35:11] robotic space. And the Optimus robots on [35:13] the way and you've got Boston Dynamics [35:15] producing their robots. And if Elon's [35:17] $20,000 Optimus robot comes out, I will [35:19] be able to touch my AI. My my AI will [35:21] exist. [35:22] >> have PMS and they won't love you and [35:24] then be really irritated with you. Which [35:27] which will decrease [35:29] cognitive load. Right? Having to manage [35:34] love [35:36] and manage moods and ups and downs. That [35:38] increases cognitive load. That increases [35:42] our ability for our brain to develop. If [35:46] I'm with the perfect partner that never [35:49] is irritated with me and I never have to [35:52] change my behavior to be better, that's [35:55] probably not good for my brain. [35:57] >> Mhm. [35:57] >> The way that the brain matures [36:01] is is through struggling, number one. [36:04] You have to learn from your mistakes. [36:06] The brain was designed for that. That's [36:07] what the brain is really good at. I mean [36:10] being able to adapt and to be able to [36:12] adjust to new situations. [36:16] Uh [36:17] that that what the [36:18] AGI is by the way. [36:21] Artificial general intelligence is is [36:23] that adaptability to different context, [36:25] different places, different cultures. [36:27] >> So, AI and ChatGPT is removing the [36:30] struggle. [36:31] >> No, no, it's it's it's it's There's [36:33] There's There's [36:34] >> Well, Annie didn't look like I mean, [36:35] Annie looked like she was cooperative. [36:38] >> But even when it comes to just doing my [36:40] day-to-day tasks, it's it's removing the [36:42] struggle of me having to think [36:43] critically. In fact, when you're [36:44] speaking, I can just type what you say [36:46] into ChatGPT, and it can spit out [36:49] another question to ask you. So, as an [36:50] interviewer, I could theoretically sit [36:52] here all day and just [36:53] defer my my questions. [36:54] >> You develop grit. You develop grit [36:57] through struggle. [36:59] >> That's right. [37:00] >> And learning. [37:02] Long-term potentiation when you learn [37:04] something new, it's hard because it's [37:07] new. [37:08] >> And what are your generally what are [37:09] your biggest concerns with artificial [37:11] intelligence, and how do we navigate [37:13] those concerns? [37:14] Is it You talked to me [37:15] >> It's out of the box. [37:17] So, I think we have to talk about it. We [37:19] have to legislate it. [37:21] Um we have to study it. Why do we keep [37:25] releasing things [37:28] that are so sexy that we don't study the [37:31] impact? We have the sickest [37:35] young generation [37:38] in the world's history. 58% of teenage [37:41] girls report being persistently sad. 32% [37:45] have thought of killing themselves. 24% [37:48] have planned to kill themselves, and 13% [37:51] have tried to kill themselves. The CDC [37:53] study. We have the sickest generation [37:56] history because we've unleashed [37:59] cell phones, social media [38:03] without any neuroscience study. If we [38:06] don't learn it, and I think AI is much [38:09] more dangerous. [38:10] Has the potential to be much more [38:12] dangerous because it's way sexier. [38:17] >> I think we are probably grossly [38:20] underestimating the impact it's going to [38:22] have. I think just like social media [38:24] where we thought the promise was that it [38:25] was going to connect us, [38:27] it's um [38:29] it's it's we're guinea pigs in an [38:30] experiment where we're going to find out [38:32] the results of the experiment probably [38:34] 20, 30 years down the line. I tend to [38:36] think people will do in the near term [38:37] what's easiest, fastest, and cheapest [38:39] and what gives them the nearest the the [38:40] short-term advantage. So, with that in [38:43] mind, I think okay, I think people's [38:45] ability to think critically is probably [38:46] going to erode to some degree. If I had [38:48] to counter my own argument, I'd say [38:51] um [38:52] am I I'm probably learning more now that [38:55] I use ChatGPT. I'm learning more [38:56] information, but I'm probably [38:59] losing my ability to think critically. [39:02] And I think they're two very different [39:03] things. Like in school, I memorized [39:06] German to pass the exam. I can't speak [39:08] German now because I just memorized the [39:10] words I needed to pass the exam. I [39:12] didn't understand German. [39:13] And I think that's kind of what's [39:14] happening. I might be able to [39:15] regurgitate things, but whether I [39:17] understand them, I think is question [39:19] question mark. And actually, as someone [39:20] who's built my my life, my fortunes, [39:22] everything, my business is based on my [39:23] ability to [39:24] innovate and think critically about the [39:26] problem and then come up with a slightly [39:28] novel solution which [39:30] learns from, you know, different first [39:31] principles to create something new. I'm [39:33] concerned that my own ChatGPT usage is [39:36] going to make me less effective. [39:39] And I'm wondering if I should put some [39:41] rules in place for myself [39:44] so that there's self- [39:44] >> Self-regulation. [39:45] >> Yeah, self-regulation. I have to do the [39:47] same with social media. On my phone, I [39:48] turn off my notifications. I have so [39:50] many things on my social media apps to [39:52] stop me using them. I don't even, [39:54] frankly, I don't even open the TikTok [39:56] app. I don't think it's even on my phone [39:58] because I think the algorithm is that [39:59] addictive. It's not to say that we don't [40:01] we don't post. My team doesn't post. [40:03] But I don't I just think yeah. [40:06] And uh [40:07] >> Well, I wrote down a couple of thoughts [40:09] I had. [40:10] >> Um [40:11] use it to amplify, not replace thinking. [40:15] >> Okay. [40:16] >> Um [40:17] alternate AI assisted with brain only [40:20] tasks. [40:22] Engage in deep learning problem solving [40:25] and memorization, so you can actually [40:27] ask AI to test you. [40:31] >> Mhm. [40:31] >> So, you're interacting with it. You're [40:34] not using it as a replacement for your [40:38] brain. And I think [40:41] just like you said, it's here and it's [40:44] going to get bigger. I think the [40:45] unintended consequences is not going to [40:47] be 20 or 30 years. I think it's going to [40:49] be five. I think like everything is [40:53] accelerated. [40:55] And I think we have to be studying kids [40:59] and the impact it has. This is just like [41:02] they did with the MIT study. [41:05] These are kids who didn't use it at all. [41:07] These are kids who use search. These are [41:09] kids that used AI. And when we see [41:13] information like this, we act on it. And [41:17] we educate kids about it. I think that's [41:22] If you can engage them, that's what I [41:24] found with my work with teenagers. If [41:27] you can get them [41:29] to really understand, okay, what is it [41:31] you really want? And do you want to give [41:35] away part of your mind share for people [41:39] who are making money on you? And I think [41:41] if you engage the There's a great new [41:45] article on revenge and the brain and how [41:49] revenge works on the nucleus accumbens, [41:53] part of the basal ganglia, that people [41:55] actually get addicted to revenge. But if [41:58] you can get them engaged in the truth, [42:02] that these companies are making money [42:04] the more they steal your mind, it'll [42:08] upset them enough that they'll begin to [42:10] supervise it. [42:11] >> I like the idea of asking Chat GPT to [42:16] give me negative feedback. I'll bet [42:17] you've done that, right? [42:18] >> Yeah, all the time. So, I'll say this is [42:20] this is my I've written this memo. I did [42:22] it yesterday. [42:23] I wrote a a two-page memo about me [42:27] wanting to introduce a new role into my [42:29] into my company. [42:31] And I went I did everything. I did like [42:32] how we'd measure if this was a success, [42:34] the background context, the person, how [42:36] the organization would be structured, [42:37] the impact they'd have, how who they'd [42:38] report to. And then I put it into all [42:41] three of the Chat GPT models I use, [42:43] Gemini, Chat GPT, and Grok, and said, [42:48] "Critique my work and tell me how I [42:49] could have written this better, [42:50] pretending that you're a top consultant [42:52] from Boston Consulting Group." [42:54] And it went through and it gave me a big [42:56] analysis of how I can make it better, [42:57] and I read what it said, and it said, I [42:58] remember it said, um, actually that was [43:00] the thing that said, "You need to [43:01] include [43:02] a financial forecasts about the impact. [43:05] You need to think about who's going to [43:07] report to whom more clearly, etc., etc." [43:09] So, I went back into my memo and I added [43:11] those things in. But, I have to I, you [43:12] know, [43:13] >> So, you're interacting with it. [43:16] >> Because I'm because I'm scared. Most [43:17] people don't do that. [43:19] I don't think I would do what I did. I [43:21] don't think I would have spent 4 hours [43:22] writing that. I could have within 30 [43:24] seconds said, "Hey, can you write me a [43:26] this job description?" And you it knows [43:29] my company now because Chat GPT has [43:31] memory. "Write me a job description for [43:32] this role. I want them to start this new [43:34] department for me." And I could have [43:36] saved myself 3 and 1/2 hours. The only [43:38] reason [43:38] >> why you're the CEO of your company. [43:41] >> Yeah, exactly. The the reason why I [43:42] didn't take the 30-second route is [43:45] because I reflect on being 23 years old [43:48] and the profound impact that writing and [43:51] simplifying had on my life. [43:53] Had I not spent 5 years writing every [43:56] single day and simplifying it into 140 [43:58] characters so I could tweet it, I [44:00] wouldn't have been religiously attached [44:01] to this idea. [44:02] >> you know what part of your brain was you [44:04] were taking advantage of? [44:06] It it was the basal ganglia. [44:09] That's repetitive. It needs practice, [44:11] practice, practice. And And once you put [44:13] that foundation in, then you become much [44:16] better cognitively. The cognitive part [44:18] There's two big learning systems. [44:20] And they have to work together. And And [44:22] so maybe I think I think that the real [44:24] problem with children [44:26] is that we our schools now is getting [44:29] away with rote learning. They call it [44:31] rote as if it's something bad. No, [44:33] that's practice. That [44:34] You know, you need to have a foundation. [44:36] You have to memorize things. [44:38] It It math, reading, and so forth to [44:41] become fluent. You need to be fluent. [44:44] And And it's the basal ganglia. And And [44:46] that's There's no basal ganglia in these [44:49] uh chatbots. [44:50] >> One of the One of the things I've [44:51] noticed just in the short term is I'm [44:53] getting lazier and lazier with spelling. [44:55] Because ChatGPT and these large language [44:57] models are so It's not spell check like [44:59] we used to have on on Word documents. [45:01] They are so good at knowing what word I [45:03] meant. So now I've I've started to learn [45:06] that I literally only need to half spell [45:08] a word. I literally mean if it was a if [45:11] it was a 12-letter word, I need to get [45:14] six letters right. [45:16] And it will know. [45:16] >> And you don't And grammar it'll fix your [45:18] grammar. [45:19] >> Yeah, it knows exactly what it means. So [45:20] like I've got ChatGPT open here. I'm [45:22] going to butcher everything. I'm going [45:23] to not look and I'm just going to say um [45:25] I'm going to say [45:33] Okay. [45:34] So that is what I wrote. I butchered it. [45:37] I tried to type with my eyes closed [45:39] looking away on my iPad. Tell me [45:40] everything I know about Daniel Amen. I [45:42] spelled the words [45:44] pretty much all wrong. And it says, [45:46] "Here's a full profile of Dr. Daniel [45:48] Amen." And I spelled every single word [45:50] wrong. [45:50] >> Wow. [45:51] >> And I didn't spell just spell them [45:52] nearly wrong. I spelled them [45:53] horrifically wrong. And it So what in [45:56] the future I come back to ChatGPT and [45:58] go, "I only need to half spell." [46:00] I don't need to spell anymore. [46:02] Just need to half spell. [46:04] >> I learned to spell [46:06] with phonics. [46:07] The sounds of letters. I suspect you [46:09] would did too. In our generation, that [46:11] was the way that was taught. You can't [46:14] teach phonics in California schools and [46:16] you haven't for the the generation. [46:18] >> Which changes their brain. [46:20] >> It completely changes their brain and [46:21] now they can't spell. I think it's a lot [46:24] of it is the fact that we're no longer [46:25] using the learning that we did which is [46:28] by rote, by memorizing stuff, by [46:31] repeating stuff, by doing problems over [46:34] and over and over again until it's [46:35] automatic. [46:36] >> You've written so much and you're well [46:39] known for being someone that teaches [46:41] people how to learn better. If you were [46:43] trying to help me learn better [46:47] based on everything you know about the [46:48] brain [46:49] what advice would you give me? I'm [46:51] someone that sits here with these [46:52] experts all day every day consuming all [46:54] of this information. Not all of this [46:55] sticks. [46:57] >> and this is something we've known for a [46:58] hundred years. [46:59] And that is if you want to remember [47:01] long-term, you should you should you [47:03] should rehearse [47:05] at intervals. [47:08] Okay, in other words, if you find an [47:09] amount of time to study something, you [47:10] shouldn't spend all that time [47:13] in one go. But if you if you spend [47:16] you know, you learn something and then [47:18] you come back the next day and you know, [47:20] rehearse it or or even better, come back [47:22] the next week and rehearse it [47:25] that spacing is something that helps the [47:28] brain solidify those memories. It's it's [47:31] called the spacing effect. Goes back to [47:32] Ebbinghaus. You go to schools [47:35] they don't teach that. They they don't. [47:36] I mean, this is the one of the most [47:37] basic facts that we've we've known about [47:39] it but [47:40] it covers every single kind of learning. [47:43] You know, cognitive learning, even even [47:45] >> don't teach us how to learn. Which we'd [47:48] think that's the first thing they should [47:50] teach us is how to love and care for our [47:53] brains and then how to learn. [47:56] >> Absolutely. Absolutely. I mean, in other [47:58] in other words, there's and you're [48:00] referring to I have a a massive open [48:02] online course, a MOOC with Barbara [48:04] Oakley on learning how to learn. It's [48:06] it's fabulously uh [48:08] popular. It's the 6 million people have [48:10] taken the course. A bunch of 50 [48:13] 10-minute segments, but the one that's [48:14] the most popular is how to avoid [48:16] procrastination. [48:19] >> And what's the answer? [48:20] >> The reason why you procrastinate is that [48:22] there's some [48:23] mental block or some energy barrier, [48:26] right? And so what you have got to do [48:28] is [48:29] get over that. And you don't do it by [48:31] just running over it. What you have to [48:32] do is say, "I'm going to spend 20 [48:34] minutes today getting started with that [48:36] task. I know it's going to take me a [48:37] long time." [48:39] I have a timer [48:40] and I start thinking about it [48:43] and I get a little bit into it, maybe [48:44] make a list. [48:46] Bang, that's the end. Okay, it's great. [48:48] It's 20 minutes. [48:50] Now, here's what happens. [48:52] You go to sleep. [48:53] Your brain is now working on that list [48:56] and you come back the next day and spend [48:58] another 20 minutes. [49:00] And you do it in small segments. You [49:01] don't want to do it all at once. And [49:02] it's just like the same thing with the [49:04] spacing effect is your your brain needs [49:07] time, your subconscious needs time to [49:09] work on things. [49:11] And so by putting in a little bit, it'll [49:13] work on it overnight and and now you [49:15] when you come the next day, you'll be [49:16] ready for the next you know, you'll be [49:19] able to build on what you've done in [49:20] your brain. [49:21] >> Is this why people say, "I'm going to [49:22] sleep on it" [49:24] when they you know, they've got [49:25] difficult problems? [49:26] >> sayings actually have meaning. It's [49:28] absolutely right, yeah. [49:29] >> Because the brain there's something [49:31] about spacing out. [49:32] >> It it's spacing, but this mem- memory [49:34] consolidation I'm talking about is is [49:36] very it's very interesting. It's [49:38] something I've I've actually worked a [49:39] lot on. And it what was happening is you [49:41] have to take the new experience and [49:43] integrate it into your old long-term [49:44] memory. [49:45] And that has to be done in a way that [49:48] doesn't interfere with there. [49:50] And also you get a chance to sort out, [49:53] you know, what's relevant, what's [49:54] important. [49:56] I know when I wake up in the morning, [49:57] things that were very muddled and things [49:59] become clearer because I think it's it's [50:01] it's eliminated a lot of things that are [50:03] irrelevant or not needed. And so you now [50:06] can see what's important. [50:07] >> So what are the things that we [50:09] do where we think we're learning [50:11] something, but they're actually not [50:12] working? [50:15] You know, cuz I'm I'm you know [50:17] I might be preparing for this podcast [50:19] today. I've got 20 pages of research [50:21] that I've pulled together. And I might [50:23] tell myself that the way to for me to [50:24] really learn that so that I don't have [50:26] to look at the research is by just [50:28] re-reading it over and over again. [50:30] >> What you should have done is not just [50:31] read it over and over again. In fact, [50:32] one of the things that we say you know, [50:34] this is a standard thing is that [50:36] students they get a mental block and [50:38] they keep banging your head against the [50:39] wall. I can't understand it. I can't [50:41] understand it. [50:42] What you the right thing to do is once [50:44] you get to that point is just get up and [50:46] start walking around doing something, [50:48] you know, cooking, [50:49] gardening, whatever it is. [50:51] Let your subconscious work on it. You [50:53] know, the brain saturates very very [50:55] quick. So having breaks at meetings is [50:59] you might think is is a waste of time, [51:01] but actually it's the most important [51:03] thing you can add to [51:04] a long string of talks is have breaks [51:06] between the talks so that you can your [51:08] brain can work on it. [51:10] And my favorite meeting actually is a [51:12] ski meeting. [51:14] And the idea is that you go to a ski [51:16] resort [51:18] and what you do is you in the morning [51:20] you have a couple of hours of lectures [51:22] and now you go skiing and now it turns [51:24] out your brain is working on what you [51:26] heard [51:27] and then when you come down [51:30] in the evening you have another couple [51:31] of hours, but now your brain is [51:33] refreshed and so it's able to take in [51:35] the new information and integrate it. [51:38] And then you go to sleep and that you [51:39] know, it's like kneading bread. You have [51:41] to go back and forth, back and forth, [51:42] back and forth. And so I found those at [51:45] the efficient in terms of learning new [51:47] things and [51:49] uh being able to think about it and mull [51:51] over it during the time of the meeting [51:54] as opposed to at the end of the meeting. [51:56] >> Every single one of you watching this [51:58] right now has something to offer, [51:59] whether it's knowledge or skills or [52:00] experience, and that means you have [52:02] value. Stan Store, the platform I co-own [52:04] who are one of the sponsors of this [52:05] podcast, turns your knowledge into a [52:08] business through one single click. You [52:10] can sell digital products, coaching, [52:12] communities, and you don't need any [52:14] coding experience either, just the drive [52:16] to start. This is a business I really [52:18] believe in, and already $300 [52:21] has been earned by creators, coaches, [52:23] and entrepreneurs just like you have the [52:24] potential to be on Stan Store. These are [52:26] people who didn't wait, who heard me [52:28] saying things like this, and instead of [52:29] procrastinating started building, then [52:31] launched something, and now they're [52:33] getting paid to do it. Stan is [52:34] incredibly simple and incredibly easy, [52:36] and you can link it with a Shopify store [52:38] that you're already using if you want [52:39] to. I'm on it, and so is my girlfriend [52:41] and many of my team. So, if you want to [52:43] join, start by launching your own [52:44] business with a free 30-day trial. Visit [52:47] stevenbartlett.stan.store [52:49] and get your setup within minutes. [52:51] I've met and invested in many [52:54] early-stage founders over the years, [52:55] probably about 50 or 60, ones like Ross [52:57] from Cadence and Marissa from Perfect [52:59] Ted. And one thing they all know is that [53:01] having a digitally fluent business is [53:03] crucial, but it isn't always easy [53:05] getting your business or team to that [53:07] point. Through my ongoing partnership [53:09] with Vodafone Business, I've seen the [53:10] work that they're doing supporting [53:12] founders and small businesses to become [53:14] digitally savvy. They know how much [53:16] small business owners value advice from [53:18] those who've been there and done it [53:19] before. So, they've just launched a new [53:21] content series to share experiences from [53:23] like-minded founders. It's called [53:25] business.connected, [53:27] part of Vodafone's support program. It's [53:29] a collection of resources designed to [53:31] support businesses with free digital [53:33] skills. So, if you've been trying to [53:35] figure out AI, marketing, e-commerce, or [53:37] just how to scale smarter, there's [53:38] advice and insights throughout this [53:40] series from those who've have done it [53:41] before. A bunch of different founders [53:43] and experts who have been there and done [53:45] it. I highly recommend you go and check [53:47] it out. Just search Vodafone [53:49] business.connected [53:50] on YouTube or follow the link in the [53:52] description below. So, let's talk about [53:54] other things outside of AI that we can [53:56] do to have good, healthy brains. Based [53:59] on everything you know about how the [54:00] brain works. [54:02] Um let's start with children. I'm hoping [54:05] to be a father at some point in the next [54:08] next couple of months or years or [54:09] whatever God God grants me a child. Um [54:13] what should I be thinking about with my [54:14] child's brain to make sure it's healthy? [54:16] >> To get your body and your partner's body [54:21] as healthy as you can before you [54:23] conceive. Cuz there's a concept [54:28] I like called brain reserve. Brain [54:31] reserve is the extra function tissue you [54:33] have to deal with whatever stress comes [54:36] your way. And it starts from the health [54:39] of the egg and the health of the sperm [54:43] that create the baby. So, there things [54:47] you guys can do now that would be really [54:51] helpful. And then once your partner is [54:54] pregnant, you want to not put her under [54:57] a lot of stress [54:59] because her body's health while she's [55:04] creating the baby. [55:06] I mean, the babies eat the brain starts [55:09] to develop I think day 21. So, even [55:12] before you know she's pregnant, the [55:14] baby's brain is developing. So, [55:19] knowing you, intentional, purposeful, [55:22] it's like, let's live as cleanly as we [55:26] can. I think that gives the baby a head [55:29] start. And then you think about what to [55:32] feed the baby. You think about what the [55:36] baby's exposed to and what the baby [55:39] needs most is mom's and your time and [55:44] eye contact and cuddling and singing and [55:50] it's like those are [55:51] >> Touching is really important. Well, but [55:53] there's another fact there was a study [55:55] that was done on [55:56] the impact of how many words are spoken. [55:59] You know, there when a a baby and a [56:01] child even when you know, baby doesn't [56:04] speak, you know, until like 18 months. [56:06] Uh but it turns out that the words that [56:08] you are talking to the baby are going [56:10] into the brain and have an impact. Every [56:13] and and and in families that don't talk, [56:17] they do worse at school. [56:20] Unfortunately, a lot of poor families uh [56:22] but but that's really important is that [56:24] they they they have they're exposed to [56:27] language early and abundantly. [56:30] >> And you model. I mean, it's the one big [56:32] thing. Whatever you want the baby to [56:35] grow into, every day you are modeling [56:38] health or you are modeling illness just [56:42] by what you do, by what you say, by how [56:44] you treat the baby's mother. [56:47] Um [56:48] I have a book called Raising Mentally [56:50] Strong Kids, which I'm very happy about. [56:52] Um [56:54] and it starts with what kind of dad do I [56:57] want to be and what kind of child do I [56:58] want to raise. And [57:02] bonding, you want your child to pick [57:03] your values. [57:06] Then bonding is time, actual physical [57:09] time. [57:11] And listening. Like being And that's [57:14] what AI does, I think. It'll actually [57:17] listen without interrupting you [57:21] and try to reflect back what you're [57:24] hearing and then give you some positive [57:26] input. Too often, because of screens, [57:29] parents aren't listening, their heads [57:31] are in their phones, and everybody's [57:35] distracted. You see it whenever you go [57:36] to a restaurant. It's like everybody's [57:38] on their phone, and nobody's looking at [57:40] each other. [57:41] >> Are we raising mentally weak kids [57:43] because we're there's a culture now of [57:45] like helping them too much? [57:48] Doing too much for them. [57:49] >> This generation is the most in trouble [57:52] in history. [57:53] And we have to really ask ourselves why. [57:58] From the food we feed them to the [58:00] devices they look at, to the negative [58:02] news, the polarization [58:05] of the news. It's that sort of chronic [58:08] cortisol. And then the separation, [58:12] oh, you voted this way or you voted that [58:15] way. Saw something TV this morning if [58:17] somebody [58:18] voted one way, well, you shouldn't spend [58:20] time with them. I'm like, we're already [58:22] so lonely that now you're going to cut [58:24] off 50% of the population. It's like [58:27] it's just such stupidity. [58:30] >> Do you think much about the the impact [58:31] that religion and having a belief in [58:34] some kind of transcendent thing has on [58:38] the brain and psychology and psychiatry [58:40] generally? [58:41] >> you don't believe in God, you're three [58:43] times the risk of depression. [58:47] Yeah, God in [58:48] different ways. [58:50] >> transcendent or [58:51] >> Yes. If you believe you're here by Just [58:55] think about it with me. If you believe [58:57] you're just here by random chance, that [59:00] life really was not created and has no [59:04] meaning, there's existential [59:07] nothingness to that. [59:10] As opposed to, oh, no, I'm created in a [59:13] special way to do something purposeful [59:17] on Earth. There's purposeful people live [59:21] longer, they're happier. [59:23] Now, whatever [59:25] version [59:26] you believe, [59:28] to not believe is hard for the brain. [59:32] And there's an interesting study on [59:35] believers versus non-believers. And you [59:38] know, many scientists would go, well, [59:40] they'll have smaller brains if they're a [59:42] believer. They actually bigger temporal [59:44] lobes. And temporal lobes, underneath [59:47] your temples and behind your eyes. [59:50] Right here. [59:52] Um [59:53] that's where it's called the God area [59:56] because um that's where people think [60:00] they experience them. [60:01] >> And and if you have a seizure in the [60:03] temporal lobe, you you have transcendent [60:05] experiences like you're [60:08] um [60:09] uh uh you know, in the presence of God. [60:11] >> And they think maybe the Apostle Paul on [60:14] the road to Damascus had a seizure and [60:18] saw God. There's actually a [60:20] uh researcher in Canada, uh Laurentian [60:23] University, Michael Persinger. So, he [60:26] would stimulate the outside [60:29] He He would do it all over the brain, [60:30] but what he found he stimulated the [60:32] outside of the right temporal lobe, that [60:35] people would get a sensed presence. They [60:37] would actually feel the presence of God [60:42] in the room. So, does that being the [60:44] brain makes up God, or does that mean [60:47] there's a way for God to communicate [60:50] with us? I actually did a study on [60:53] prayer. It was so interesting. You know, [60:55] I pray for you, uh prophecy, something [60:59] called speaking in tongues. Um it was [61:01] fascinating. Speaking in tongues is [61:04] channeling. It means you're channeling [61:06] the Holy Spirit. And the hypothesis was [61:09] you'd have to drop your frontal lobes, [61:11] which is exactly what happened in 60% [61:15] of our patients. And one, basal ganglia [61:18] skyrocketed, just like got hit with [61:20] cocaine. Cuz that's where cocaine works [61:23] in the basal ganglia. [61:25] Uh [61:26] so interesting. [61:27] >> If you had to create a brain-healthy [61:29] nation and I made you president of the [61:30] United States [61:32] for 1 month, [61:33] and you had to put in place executive [61:35] orders [61:37] that would create a brain-healthy [61:38] nation, what executive orders would you [61:40] immediately sign? [61:42] >> One question. [61:44] Get all of the departments ask [61:47] themselves, what we're doing is this [61:49] good for our brains or bad for it? [61:51] And so I that's the campaign. I mean, I [61:54] realize I've been doing this a very long [61:56] time. If I can just get people to answer [61:59] that one question with information [62:03] and love, love of themselves, love of [62:05] their families, love of their country. [62:08] Is this as what we're doing good for our [62:11] brains or bad for it? [62:13] >> By far, the best drug you can take for [62:16] your brain, and not just your brain, but [62:18] your entire body is exercise. [62:22] In other words, [62:24] exercise [62:25] you pump the blood and your brain gets [62:28] you know, [62:29] a lot of nutrients and everything. [62:32] It helps your heart, it helps your [62:33] immune system. People don't realize how [62:36] important that is. And we're not talking [62:38] about being an athlete, we're just [62:39] talking about walking if you're older. [62:41] Walking is perfectly good exercise. [62:44] And and you know, children now, I you [62:46] know, they're they're not getting enough [62:47] exercise. [62:48] >> they're on devices. [62:49] >> Yeah. [62:49] >> And so I have a model, if you want to [62:52] keep your brain healthy or rescue it, [62:54] you have to prevent or treat the 11 [62:56] major risk factors. And we've talked [62:58] about them before. [63:00] Exercise helps you with every single [63:03] one. So like, it's called bright minds. [63:06] So B is for blood flow, increases blood [63:08] flow. Retirement and aging, it decreases [63:11] your age. I is inflammation, it's [63:12] anti-inflammatory. G is genetics, it [63:15] helps turn on health-promoting [63:18] genes. H is head trauma. If you keep [63:22] walking, you're less likely to fall when [63:24] you're older, right? T is toxins, sweat [63:27] detoxifies you. M is mental health, [63:32] exercise boosts dopamine, but it also [63:34] boosts serotonin. So, it's like that [63:36] perfect balance or in your brain. [63:38] >> Breathing, how we breathe. [63:40] Does that have an impact on brain [63:42] health? [63:43] >> So, you can almost immediately improve [63:46] heart rate variability, which is a sign [63:49] of heart health, but also goes to mental [63:52] health by breathing [63:54] in a certain helpful way. And I call it [63:58] the 15-second breath. So, 4 seconds in, [64:02] big breath, hold it for a second and a [64:04] half, pause just a little bit, [64:06] 8 seconds out, hold it for a second and [64:10] a half. So, if you take twice as long to [64:13] breathe out as you breathe in, it [64:16] increases something called [64:17] parasympathetic tone. [64:19] And it just calms you down almost [64:21] immediately. So, if you're having panic [64:23] attacks, yes, you can take Xanax, but [64:25] there's so many problems with that later [64:27] on. Or you can just learn how to [64:30] breathe. We call it diaphragmatic, so [64:32] breathe mostly with your belly, taking [64:35] twice as long to breathe out as you [64:37] breathe in. [64:37] >> Chewing. [64:38] Uh there's a [64:39] piece here that says it stimulates [64:41] hippocampal activity and may slow [64:43] cognitive decline. Reducing chewing has [64:45] been linked to impaired learning in [64:47] animal studies. [64:48] >> And fast food decreases chewing because [64:51] it's fast, so they take most of the [64:52] fiber out so you can chew it faster or [64:54] you can swallow it faster. [64:56] >> Things in the bad for your brain list. [64:58] Overuse of GPS navigation app, which [65:00] weakens the hippocampus by outsourcing [65:02] spatial memory long-term. This can lead [65:04] to atrophy in areas associated with [65:06] memory and navigation. [65:07] >> And people are diagnosed with [65:08] Alzheimer's disease later in life [65:11] because of Siri. [65:14] Because I used to like when I started as [65:16] young psychiatrist, somebody get lost in [65:19] a city they'd lived in for 30 years and [65:22] their family would call me upset. [65:25] And I'm like, okay, this person's headed [65:27] toward dementia. Now that person goes, [65:30] take me home. Do you think it's going to [65:32] have we're going to have an epigenetic [65:34] effect [65:36] of not reading maps? That if Steven, now [65:41] he uses his phone to get from A to B, do [65:44] you think that's going to affect [65:45] Steven's son or daughter [65:48] because dad [65:51] didn't have [65:52] >> Wow, okay, that that never occurred to [65:54] me that you could pass on something like [65:57] that. By the way, I I think it has to be [65:59] physiological. Stress, for example, [66:01] could be probably passed on. You [66:02] mentioned this. You you pointed out [66:04] during pregnancy you you want to prevent [66:06] stress and and crisis, right? [66:08] Do Do you know about [66:09] >> that study with mice where they made [66:11] them afraid of the scent of cherry [66:13] blossoms from memory and [66:17] so whenever the mice smelled [66:20] cherry blossoms, they would shock them. [66:23] Mildly. So the mice are now afraid of [66:25] the scent of cherry blossoms. [66:27] Their babies were afraid of the scent of [66:29] cherry blossoms. Their grand babies were [66:31] afraid of the scent of cherry blossoms. [66:33] So [66:33] >> that's the olfactory system. Olfactory [66:35] system is very interesting because it [66:37] goes directly to the hippocampus. [66:39] But there might be an evolutionary [66:41] advantage because if if there's [66:44] something in the environment that you [66:45] shouldn't eat or you know that smells a [66:48] particular way, passing that on is is a [66:51] very efficient instead of having to have [66:53] experience that yourself, you know, [66:54] trial and error because it might if a [66:56] poison is, right? It might kill you. But [66:58] if if you've your parents [67:01] had that bad experience and passed it on [67:03] to you shouldn't go to the something [67:05] that smells in a particular way, that [67:06] makes sense. [67:07] >> The other thing that's bad for the [67:08] brain, which is unexpected, is you said [67:11] at the start, artificial sweetness. [67:14] Now, I didn't think I thought artificial [67:16] sweeteners were fine. [67:17] >> They're not fine, and they're not free. [67:19] So, I used to drink diet soda like it [67:21] was my best friend, cuz I thought it was [67:23] free. [67:24] Um then I had arthritis when I was 35, [67:28] and one of my patients said she stopped [67:31] aspartame and her arthritis went away. [67:33] And I'm like cuz I was drinking like I [67:35] don't know. [67:36] A lot of diet soda, and so I stopped, [67:40] and my arthritis went away. And I'm [67:41] like, no. And so I did it again, and it [67:44] came back, and I'm like, okay. [67:46] And [67:48] artificial sweeteners can change the [67:50] microbiome. So, we haven't talked about [67:53] that, but you have these 100 trillion [67:55] bugs in your gut that make [67:57] neurotransmitters, digest your food, and [68:00] especially sucralose or Splenda has been [68:03] found to decrease the good bacteria in [68:05] your gut, which then has a negative [68:07] impact on brain function. [68:10] >> It is aspartame, as you mentioned. [68:12] >> And aspartame that I mentioned that can [68:14] have a generational impact. So, is it [68:16] possible it's really not social media, [68:19] it's that we've had aspartame in our [68:22] food [68:23] for decades. And I think it's [68:26] all of these things that just sort of [68:28] are additive, and we should just always [68:31] think that that one question. Is this [68:33] good for my brain or bad? For you [68:35] mention broccoli, probably that's good [68:37] for your brain. [68:40] Cheeseburger, probably not, but why [68:42] don't you take the burger, and if you [68:44] could make it grass-fed, that would be [68:46] better, and put it in a salad, and then [68:48] that would be good for your brain. [68:50] >> What about chronic background noise? [68:53] We don't think much about the impact [68:54] noise has, but [68:56] >> I used to live, um my house was three [69:00] houses from the freeway. And if you just [69:03] go there, it's like, "My god, it's so [69:06] loud here." I never heard the freeway [69:09] because my brain just learned to tune it [69:11] out. [69:11] >> By the way, was that good or bad for [69:13] you? [69:14] >> That I was able to [69:15] >> had you adapted and and were no longer [69:17] sensitive to it. I I I think that [69:19] actually was probably not good [69:22] for for various reasons because it what [69:25] it really means is that you you're [69:27] you're specializing for that [69:28] environment. And your brain's going to [69:30] be different when you go someplace. [69:33] So but so here's a here's another [69:34] example. [69:35] >> stressful. [69:37] Right? [69:37] >> Yes, chronically stressful, but my brain [69:40] is very good at tuning out. [69:41] >> But in the background in the In other [69:42] words, yeah, in other words, your your [69:43] your your brain is reacting to it even [69:46] though you're not aware of it, you know. [69:47] >> So subtly increases And I had five [69:48] sisters, which makes it even worse. [69:50] >> Subtly increases cortisol and impairs [69:52] working memory and attention regulation, [69:54] especially in children and older adults [69:55] to be chronically exposed to background [69:57] noise like traffic or the low-level hum [69:59] of the city. [70:00] Yeah, that's right. That's absolutely [70:02] right. [70:03] >> Make sure you keep what I'm about to say [70:04] to yourself. I'm inviting 10,000 of you [70:07] to come even deeper into the Diary of a [70:09] CEO. Welcome to my inner circle. This is [70:12] a brand new private community that I'm [70:14] launching to the world. We have so many [70:16] incredible things that happen that you [70:17] are never shown. We have the briefs that [70:20] are on my iPad when I'm recording the [70:21] conversation. We have clips we've never [70:23] released. We have behind-the-scenes [70:25] conversations with the guest and also [70:26] the episodes that we've never ever [70:29] released and so much more. In the [70:32] circle, you'll have direct access to me. [70:34] You can tell us what you want this show [70:35] to be, who you want us to interview, and [70:37] the types of conversations you would [70:39] love us to have. But remember for now, [70:41] we're only inviting the first 10,000 [70:43] people that join before it closes. So if [70:46] you want to join our private closed [70:47] community, head to the link in the [70:48] description below or go to [70:49] DOACcircle.com. [70:52] I will speak to you there. [70:54] >> Many of us multitask across multiple [70:56] screens now. We're watching TV here. [70:58] We've got our phone here, we've got our [70:59] iPad here, got our computer here. And I [71:01] was reading into the science of [71:02] multitasking and it said that it trains [71:03] your brain to be distractible. [71:05] Reducing gray matter density in the [71:07] anterior cingulate. [71:10] >> Yeah, that's [71:10] >> And the insula. [71:11] >> It's [71:11] yeah, in the medial prefrontal cortex. [71:14] >> And when the insula the insula is so [71:16] interesting and I know you can talk [71:18] about I have a new study coming out on [71:19] hope. So, on 7,500 patients we gave them [71:24] a hope questionnaire. [71:26] Hope questionnaire. Hope, like how much [71:28] hope do you have that you have the [71:31] ability to make tomorrow better. And [71:34] people with low hope have lower overall [71:38] prefrontal cortex function, but the [71:41] insula [71:43] was really low and that signal was the [71:46] most statistically significant of the [71:49] group. Really. And [71:51] in some studies the insula is called the [71:53] island [71:54] >> also for depression, people who have [71:56] depression [71:57] uh have low activity in the anterior [71:59] cingulate and in fact uh [72:00] it it deep brain stimulation has been [72:03] used now for to help some people if you [72:05] stimulate that area. [72:06] >> And what our imaging research would say [72:08] is depression is like chest pain. [72:11] It's not one thing. Like nobody gets a [72:13] diagnosis of chest pain because that [72:16] would be stupid, right? It could be [72:17] heart attack, heart arrhythmia, heart [72:19] infection, gas, grief. [72:22] Depression's the same way. When you look [72:24] at it from an imaging standpoint, [72:26] sometimes their frontal lobes are too [72:27] active, sometimes they're not active [72:30] enough, sometimes it's their limbic [72:34] system that's too active. And I wrote a [72:38] book called Healing Anxiety and [72:39] Depression and like here's the seven [72:41] things I see as an imager. [72:43] >> What about ADHD? [72:45] There's obviously been a rise in ADHD or [72:46] at least people reporting of being [72:48] diagnosed with ADHD quite significant. [72:51] Can you find ADHD in the brain? Are we [72:53] causing ADHD as a function of the way [72:56] that we're living our lives, or is it [72:57] something within the brain genetically [72:58] that I could I could see? [73:00] >> So, it's both. I think Clearly, you can [73:03] see ADHD in people's families. In fact, [73:07] if I have a hyperactive, restless, [73:09] impulsive, disorganized, procrastinating [73:13] child, I'm looking at the mom and the [73:15] dad. I'm like, "So, where is this coming [73:17] from?" [73:18] But, you could also get ADHD from a head [73:21] injury, especially if it affects their [73:23] frontal lobes, which is why you [73:24] shouldn't let children hit soccer balls [73:26] with their forehead. You can also get it [73:29] from the chronic from the excessive [73:32] input, making people distracted, just [73:34] like you said. Brand new study out on [73:38] children who took medicine, right? We [73:40] always demonize ADHD medicine. But, the [73:44] kids who took medicine actually had [73:48] bigger brains in their prefrontal cortex [73:51] than kids who didn't take medicine who [73:54] had ADHD. [73:55] >> Ritalin? [73:56] >> Ritalin. [73:57] >> That's Okay, that's speed, basically. [73:59] Yeah, amphetamines. [74:00] >> It is. But, for the kids who have it, [74:04] I think withholding medicine [74:07] from a child who really has ADHD is like [74:10] withholding glasses from someone who has [74:13] trouble seeing. [74:14] And it's it's the easy thing to demonize [74:18] the drugs until you realize someone who [74:21] has ADHD [74:23] a third of them don't finish high [74:24] school. [74:26] And [74:26] we never ask the right question about [74:29] People go, "What's the side effects?" [74:30] And it's it can decrease your appetite. [74:33] And [74:35] can can have sleep problems with it. [74:38] But, they don't ask the other question. [74:39] It's what's the side effect of not [74:41] taking the medicine, or at least not [74:43] fully treating And there There other [74:45] ways to treat it besides medicine, You [74:47] know, for God's sakes, I own a [74:49] supplement company and I'm always trying [74:51] to optimize the nutrients to the brain. [74:54] Neurofeedback can help, but if you do [74:58] those things and it's not working, don't [75:00] be afraid of medicine. [75:03] >> But by the way, when I was growing up, [75:04] ADHD either didn't exist or they didn't [75:07] know about it. [75:08] Do you Do you think that there's some [75:10] link to the our diet [75:11] >> it was first described in a book named [75:13] 1910 and it's in the first version [75:17] of the DSM. [75:20] They called it minimal brain [75:21] dysfunction. [75:23] But when we were growing up, there were [75:25] one or two of these kids in our [75:27] classrooms and now there's eight to 10. [75:30] >> That's That's what I mean. Is Is that [75:31] it's [75:32] seems to that like autism, it's [75:35] seems to be proliferating. [75:37] >> Right, and part of it I think is the [75:39] food that is [75:42] much more processed. Part of it is the [75:44] screens. Part of it is the distracted [75:46] parents and part of it is the teaching. [75:50] >> You always seem to be doing new studies, [75:52] Daniel. What What new studies are you [75:54] most excited about or have you completed [75:56] since we last spoke? [75:57] >> I did one that I'm so excited about on [76:02] negativity and the brain. [76:04] And negativity is bad for your brain. So [76:10] >> you define negativity? [76:11] >> We actually give them a questionnaire. [76:14] It's a positivity negativity bias [76:17] questionnaire and people who are more [76:19] negative have less activity in their [76:23] prefrontal cortex. It's actually quite [76:26] interesting. And so [76:29] unbridled positivity is bad for you [76:33] because you need that 15%, but if you're [76:36] chronically negative, that is [76:40] bad for your brain. [76:41] >> Is there a link between being a negative [76:42] person and Alzheimer's and dementia? [76:45] >> Yes. And what's interesting cuz you [76:46] mentioned a gender difference earlier, [76:49] um [76:50] if you're depressed and you're a woman, [76:53] it doubles your risk for Alzheimer's [76:55] disease. If you're depressed and you're [76:57] a man, it quadruples [77:00] your risk. [77:01] >> Wow. [77:02] So, there was a study was done during [77:04] the COVID years, a couple years, [77:07] and it turns out that the rate of [77:10] depression like doubled in women, [77:13] but not in men. [77:15] >> During COVID? [77:16] >> During COVID. And after COVID, when the [77:18] as students came back and everybody was [77:20] back to normal, so-called normal, [77:23] the women stayed depressed. [77:26] At that high level. [77:28] Which is very is a very interesting that [77:30] it should be the women who [77:32] >> So, in [77:32] >> one study, women had 52% less serotonin [77:36] than men. So, I think is really [77:39] interesting women by and large have [77:41] double double the risk depression. Women [77:44] have double the risk of depression as [77:46] men. Their limbic systems are larger, [77:48] which is also involved in [77:50] >> probably, yeah. [77:51] >> more vulnerable in bonding. And then, [77:54] the whole COVID thing we haven't talked [77:56] about. COVID [77:58] causes inflammation in the limbic part [78:00] of the brain. I had scans of people I [78:03] was treating, [78:05] and then they got COVID, and then I [78:07] scanned them again, and you can just see [78:09] this dramatic inflammation in the brain. [78:14] >> If someone's listening now, and they [78:15] just want to they want to improve their [78:17] brain health, they want to avoid [78:18] dementia, they want to be cognitively [78:22] powerful and capable as they age. They [78:24] want to get to 80 years old, 90 years [78:26] old, 100 years old, and have a great [78:28] brain. [78:29] And you just had to and you could only [78:31] tell them to do three things. [78:33] >> Well, Terry said one, exercise. [78:35] >> Okay, exercise, I'm going to do that. [78:38] >> Start every day with today is going to [78:40] be a great day. [78:41] >> Positive day. [78:41] >> Push your brain [78:43] to look for what's right rather than [78:47] what's wrong. [78:48] >> Okay. So, I'm going to be optimistic and [78:51] grateful. [78:51] >> Omega-3 fatty acids. [78:54] And either do it with fish or do it with [78:56] a supplement. [78:56] >> Why did you include omega-3 fatty acids? [78:59] >> Because it decreases inflammation and [79:03] 25% [79:05] of the cell membranes in your brain are [79:07] made up of omega-3 fatty acids. And as a [79:10] country, we're dramatically low in them. [79:14] >> And learning. That's maybe one of the [79:15] things that's been left off the list of [79:17] top three things. [79:19] But I mean, I remember you telling me [79:20] that [79:21] how good learning was for the brain. And [79:23] even getting outside and running outside [79:25] versus running on a treadmill is more [79:27] beneficial. [79:28] >> And if you learn while you're [79:29] exercising, [79:31] what you're doing is you're getting [79:33] blood flow to the hippocampus and you're [79:35] more likely to remember it. So, [79:39] >> I heard this, yeah. I had someone tell [79:40] me that they um figured out that they [79:42] could learn better for their exams if [79:44] they did it in a sauna. [79:46] So, they kept It was a scientist that I [79:48] spoke to. She said she keeps learning [79:50] new information when she's in the sauna, [79:52] cuz she realized that when she left the [79:53] sauna and was then tested upon on it, [79:55] she was better able to uh do the exam. [79:59] And I guess that's correlating to what [80:00] you said about because in a sauna, [80:01] you're going to have a lot of blood [80:02] flow, I imagine, [80:04] to the brain? [80:04] >> Yeah. [80:05] >> There's actually a study in JAMA [80:06] Psychiatry [80:08] that one sauna [80:11] bath [80:13] helped depression. Significantly helped [80:16] depression. [80:17] And I think it's because of it's [80:20] balancing the brain. And people who do [80:22] the most saunas have the lowest risk of [80:24] Alzheimer's disease. [80:25] >> Hm. [80:27] What is the most important thing as it [80:28] relates to the subjects that we spoke [80:29] about today, AI, the brain, [80:32] neuroscience. [80:34] That you would like to say to the and [80:36] the people that are listening now. There [80:37] could be a million people listening, [80:39] there could be 20 million people [80:40] listening. If you could say one thing to [80:42] them about the brain, [80:44] AI, neuroscience, whatever you want to [80:45] say, the floor is yours. What would that [80:47] be? [80:49] Over to you first, Terry. [80:51] >> Sleep. Sleep is the time when the body [80:54] not just regenerates, but your memory is [80:57] consolidated. So, things you've [80:59] experienced during the day [81:01] are integrated into your cortex and it's [81:04] an interaction between hippocampus and [81:05] the cortex [81:07] for [81:08] you know, for for episodic memories. [81:10] And and and it's unfortunate what's [81:13] happening with children now, you know, [81:15] they're so competitive to get into [81:17] college that they're cutting back on [81:19] their sleep and it's just the wrong time [81:21] of your life. You shouldn't be cutting [81:22] it back when your brain is [81:24] is developing. So, those two things, I [81:26] would say sleep and exercise are the [81:28] most important thing for your brain. [81:29] >> The floor is yours. [81:31] What would you say to the [81:33] listeners [81:34] about all the things we've talked today? [81:35] What's your closing closing statement? [81:38] >> Well, you know, go back to what I talked [81:40] about in the beginning, which is we've [81:42] just thrown the barn door open [81:45] and let the horse bolt [81:47] out into our schools, into our [81:50] businesses, into our homes. [81:53] And before we even asked is it a gift or [81:58] is it a Trojan horse that's going to [81:59] steal [82:00] from us. We've embraced convenience [82:04] before understanding consequence. And [82:07] we've done it before with video games [82:10] and cell phones and social media and [82:12] marijuana and alcohol and [82:14] opiates and high fructose corn syrup and [82:17] aspartame. And we have to be smarter. We [82:20] have to tame this horse as it's gone [82:24] with wisdom or it's going to trample our [82:26] children. And so I think we have to be [82:29] very thoughtful and it all comes back [82:31] down to is this good for my brain or bad [82:33] for it. Is it good for our collective [82:35] brains or is it potentially bad for it? [82:39] And [82:40] just answer that question with [82:42] information and love of yourself, of [82:46] your family, of your country, [82:48] community. [82:54] Yeah, I'm more anxious than when I came [82:55] in. I don't like that. [82:59] >> It it's it's I'm just it's so upfront in [83:01] my mind for me at the moment because [83:02] I have the hindsight, the wisdom of [83:05] hindsight of all those things you [83:06] mentioned like exercise and processed [83:08] foods and social media and all these [83:10] things that we tried and they all seem [83:12] to follow a similar arc. Some kind of [83:14] new product or discovery is made. The [83:17] early phase, in that early phase people [83:19] who have an incentive for that thing to [83:21] be successful will somewhat like [83:23] gaslight you into thinking that it's [83:25] fine. [83:26] And then we get into the second phase [83:27] where we start to see sort of some [83:29] consequences then we study what's [83:30] actually happened. We figure out that [83:32] there's there was always a trade-off [83:34] and that nobody really understood the [83:36] trade-off. [83:37] And then people change their behavior. [83:39] So now when I come into these new [83:40] technologies where [83:42] the short-term benefit is really clear, [83:44] it's making me more productive, [83:47] I I pause and I go there's going to be a [83:49] trade-off here. There's always a [83:50] trade-off. What is the trade-off? And am [83:53] I comfortable and conscious of what that [83:54] trade-off is? [83:56] And if if the trade-off So I try to [83:57] figure out what the trade-off is with [83:58] things like AI and I go okay, the [83:59] trade-off is probably I'm going to be [84:00] worse at critical thinking [84:02] and then might have an impact on my [84:03] social relationships if I fall in love [84:05] with [ __ ] Annie because she's pretty [84:06] hot to be fair. [84:09] And I really value my critical thinking. [84:11] I really value my ability to um solve [84:14] problems and to articulate myself and to [84:16] write and to communicate with my loved [84:18] ones in an effective way. So what can I [84:20] do if that is the trade-off now? And one [84:23] of the things that I'm doing now feels [84:25] really counterintuitive in a world where [84:27] everybody's got these productivity gains [84:29] cuz they're using these tools, which is [84:30] to refrain. And I I wonder if one of the [84:33] great advantages of the next decade, one [84:35] of the great hedges for anyone that's [84:38] wanting to be a great critical thinker, [84:40] entrepreneur, creative, is to go left [84:43] when everyone's going right. [84:46] Which is to refrain and do it the hard [84:48] way. [84:49] And if we look at history [84:52] in these arcs that we've discovered with [84:54] food and with exercise and all these [84:55] things and dating, doing it the hard [84:57] way, like we said about the marshmallow [84:59] test and delaying the gratification, [85:01] seems to yield the greatest returns. So, [85:04] I think I'm going to do it the hard way. [85:05] >> easiest [85:07] because it won't have the side effects. [85:09] >> Yeah, the hard way. [85:10] >> Right. I want I want to feel good now [85:14] and later as opposed to now, but not [85:18] later. [85:19] >> And to be clear, this doesn't mean I'm [85:20] not going to use AI or ChatGPT. It just [85:22] means that when it matters, when the [85:24] thinking matters, I will think for [85:26] myself. [85:28] And when the communication matters, I'll [85:29] communicate for myself. [85:31] That's what I That's my conclusion. [85:32] >> You should hope that your children will [85:34] feel the same way when they grow up. [85:37] >> will model what you do. [85:40] Right? Every day you model health [85:44] or not health. [85:46] >> Mhm. [85:48] Thank you. Thank you for writing to the [85:50] Well, incredible books that I've got [85:51] around me. I'm going to link them all [85:52] for my viewers that are watching. I've [85:54] got so many of these books. Um the [85:56] incredible one that you wrote for [85:57] parents called Raising Mentally Strong [85:58] Kids. You've got your other book over [86:00] there, Change Your Brain Every Day. And [86:03] I've got this book here from from Terry, [86:05] which is The Deep Learning Revolution. [86:07] And one you wrote most recently called [86:09] ChatGPT and The Future of AI. I'm going [86:11] to link all of them below. [86:12] I'm going to link them with a little bit [86:13] of a summary of what's in them. So, if [86:15] you decide that there's anything here [86:17] that we talked about today that where [86:18] you want to dig in further, please do [86:19] dig in. And I'm also going to link [86:21] both of your link to where people can [86:23] find out more about both of you, your [86:25] websites, and more of your work in the [86:27] comments below. So please do check that [86:28] out everybody listening. We have a [86:30] closing tradition where the last guest [86:31] leaves a question for the next guest, as [86:33] you know, and they don't know who [86:34] they're leaving it for. [86:35] So I'm going to ask you both the [86:36] question. [86:38] Starting with you, Daniel. Are you [86:39] prepared [86:41] for recognition of your next [86:45] health challenge? [86:47] Will you be able to notice its onset? [86:50] And how will you address the challenge, [86:53] even if it means a major lifestyle [86:55] change or way of living? [86:59] Hm? [87:00] >> Okay. [87:02] Yes. [87:04] >> How will you address the challenge, even [87:06] if it means a major lifestyle change or [87:08] way of living? [87:10] >> Well, I'm very clear on the goals I [87:13] have, which is to be vibrant and [87:16] healthy, and not get dementia. So, if I [87:20] need to change something so that [87:22] happens, I'm like all in. [87:25] >> Are you prepared? [87:27] >> Probably not. [87:30] No, I've been blessed with good health. [87:32] I try to live a healthy [87:34] life. [87:36] But the problem is that you can't [87:39] anticipate, you know, as you get older, [87:42] what [87:43] you know, what's ahead, you know. [87:48] Like you mentioned arthritis. I'm [87:49] feeling a little bit of arthritis now. [87:51] I've been arthritis free [87:53] most of all my life. [87:55] And, you know, that's something it's [87:57] very It's to realize that it's coming, [88:00] and there's very little you can do about [88:01] it. [88:03] Is is depressing. [88:05] But on the other hand, things could [88:07] always be worse. [88:09] And sometimes that cheers you up. [88:12] But, the the the the the reality is that [88:15] there are things [88:17] in the world like COVID, you know, that [88:19] you have no control over [88:21] that may [88:23] or an accident or [88:25] Alzheimer's, you know, God forbid. [88:28] You know, that you that is who knows [88:30] what will happen, right? You you have to [88:32] live with whatever life deals with you. [88:35] You know, the time that you have you [88:36] should really spend on [88:39] uh trying to make it a [88:41] a a [88:42] healthy life, a a productive life, a you [88:46] know, satisfying life and and that's [88:47] something that we have control over, [88:49] right? [88:49] >> What are you scared of? [88:53] >> Right now, it's China. [88:57] I you know, I I I I'm not being [88:59] facetious. I I think that it's uh [89:02] it's it's it's it's what a threat that [89:04] is a societal threat. It's not I don't [89:07] think that it's going to affect me and [89:09] I've had great Chinese students inside, [89:10] but I really like I think the the [89:12] Chinese [89:14] people are different from what I see as [89:15] the the country [89:17] uh the the the what they're trying to [89:19] do, their goals that they're taking. [89:21] Uh 20 years ago, if you look at the [89:25] um [89:25] of all the technical areas in physics [89:28] and chemistry and biology and so forth, [89:30] the the 100 most important advances have [89:32] been made [89:34] in in in 20 years ago was like the [89:37] Americans had like, you know, 94 of [89:39] them. [89:41] This year [89:43] it's 74 are Chinese. [89:47] And that's because they made a huge [89:49] investment in science and STEM [89:51] research. They [89:53] They You know, they they put they poured [89:54] out a million engineers for [89:58] to to implement AI, [90:00] right? [90:02] Uh you know, they're they're doing the [90:03] right thing. We did that. Remember when [90:05] the Sputnik uh went over, the Sputnik [90:07] moment? [90:09] We made a huge investment in STEM, in [90:11] science and engineering, and and in [90:13] education. [90:14] >> What was a Sputnik moment? [90:15] >> Oh, okay, '57 when the Russians put a [90:18] satellite that went over the US over and [90:20] over again and we didn't took us years [90:23] to put up our own satellite because we [90:24] had fallen behind. [90:26] But we that investment we've been living [90:28] on [90:29] literally, you know, for the last 60 [90:32] years. [90:33] And now the the Chinese have done that [90:35] and and they're going to be advanced and [90:37] they're going to be way beyond us. [90:38] Uh you know, I [90:40] this is [90:41] You asked me, okay? That's something I'm [90:44] I was I'm very very uh [90:46] disappointed in our country that we're [90:48] we're not In fact, we're just doing the [90:49] opposite. We're tearing apart science [90:51] right now with with the present [90:52] administration. [90:53] >> What are you scared of, Daniel? [90:57] >> Losing my wife. That's the most [91:00] the thing that comes to mind. And when I [91:02] was thinking about China and my [91:05] mother-in-law was a prepper. [91:07] And [91:08] >> being someone that's preparing for the [91:09] end of the world. [91:10] >> Prepared for the end of the world. [91:12] And we were in Egypt last year and got a [91:15] call that she had cancer and we were [91:17] there for 3 days and came home. [91:20] And I kept thinking I loved her dearly. [91:23] I'm like, you prepared for the wrong [91:25] thing. [91:26] You should have prepared for cancer. [91:29] Like I think every day we should be in [91:32] the same Alzheimer's prevention program [91:34] is a cancer prevention program. It's a [91:36] heart disease prevention program. It's a [91:38] diabetes prevention program. And I'm [91:42] like, she's prepared for the wrong [91:45] thing. The thing you really want to be [91:47] prepared for is disease, right? And I [91:50] know I'm going to die. I just want to be [91:53] a vital for as long as I can be and hope [91:58] is well, I have a say [92:01] in this. [92:02] Right? Cuz I know I can accelerate [92:06] my body's decline [92:08] or I can decelerate it. And I'm going to [92:10] choose to decelerate it. [92:14] Thank you. We're done.