Panel 6 - 4: Brain Health in the Age of AI: Reclaiming Your Agency

Panel 6 - 4: Brain Health in the Age of AI: Reclaiming Your Agency — Ai4

Ai4

AI-assisted summary of Panel 6 - 4: Brain Health in the Age of AI: Reclaiming Your Agency — original source. Check the original for full context.

Source published:

Detailed notes

## Background Published on YouTube by Ai4 on August 18, 2026, this presentation features Aaron Tate, Director at the Center for BrainHealth at The University of Texas at Dallas. The session addresses the unexamined neurological consequences of widespread AI adoption, exploring whether automated thinking degrades cognitive infrastructure and offering evidence-based practices to preserve human cognitive agency.

## TL;DR Aaron Tate explains how passive cognitive delegation to AI can trigger a detrimental 'cognitive debt cycle,' leading to prefrontal cortex atrophy, conceptual homogenization, and the erosion of critical thinking. Drawing on research and longitudinal data from The BrainHealth Project, he differentiates between passive delegation and intentional cognitive amplification, where users retain primary conceptual reasoning before leveraging AI for execution. Tate outlines a four-part behavioral framework—neuroscience literacy, originating before delegating, purposeful disengagement, and reflective auditing—alongside agentic guardrails and updated SMART cognitive protocols to build sustainable neural reserves.

## Atmosphere & style The talk carries a calm, authoritative, and scientifically grounded tone, avoiding alarmism while delivering a sobering wake-up call about mental automation. Tate speaks with the measured clarity of a translational researcher, systematically moving from neural mechanisms to practical workplace strategies. It feels like a high-level strategic briefing that blends clinical neuroscience with actionable productivity design, leaving the listener introspective about their own daily mental friction and cognitive reliance.

## Key takeaways - [00:04](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=4s) The BrainHealth Project is a 10-year longitudinal study tracking adult brain health via the Brain Health Index (BHI), with clinical research showing targeted training delivers an average 7.9% cognitive improvement. - [02:51](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=171s) Executive function operates on neuroplasticity: neural pathways that are consistently challenged are reinforced, while pathways bypassed through habitual automation progressively atrophy. - [04:37](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=277s) Routinely outsourcing primary mental effort to AI initiates a 'cognitive debt cycle' characterized by declining independent critical reasoning and conceptual homogenization. - [04:37](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=277s) In an MIT Media Lab essay study, 83% of AI-reliant participants were unable to quote from or outline their own submitted work. - [04:37](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=277s) Cognitive Amplification involves doing foundational thought work before using AI for execution and tedium, whereas Cognitive Delegation outsources core higher-order synthesis directly to models. - [04:37](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=277s) As AI models become ubiquitous, competitive advantage will no longer come from access to tools, but from the personal discipline to preserve independent executive thinking. - [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) A brain-healthy AI framework relies on four habits: basic neuroscience literacy, originating ideas before delegating, disengaging to consume raw source material, and reflecting daily on cognitive effort. - [11:41](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=701s) Just as industrial physical automation led to the creation of athletic gyms, cognitive automation requires structured cognitive conditioning to maintain human analytical capacity. - [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) AI agents can be designed with intentional cognitive guardrails that prompt users to complete initial drafts and reasoning before the system offers refinements. - [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) Economic research demonstrates that investing in critical thinking and brain health generates measurable increases in individual lifetime earnings and workforce productivity. - [15:06](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=906s) Scaling validated brain health metrics and AI guardrail protocols to the broader market requires translational research partnerships with industry.

## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=4s) The Center for BrainHealth and Measuring Cognitive Potential

Artificial intelligence offers undeniable gains in productivity, efficiency, and even creative output, but realizing these benefits requires critically examining the technology's downstream consequences on brain health. The Center for BrainHealth at The University of Texas at Dallas was founded over thirty years ago by Dr. Sandra Chapman as a translational research institution. Rather than focusing exclusively on clinical deficits, the center investigates cognitive performance and human potential across all age groups, developing interventional protocols and technologies designed to actively protect and enhance brain function.

The center's flagship initiative is The BrainHealth Project, a landmark ten-year longitudinal study designed to track brain health in healthy adults at six-month intervals. The project is actively recruiting a cohort of 100,000 participants, with a designated subpopulation undergoing direct neuroimaging. Progress is quantified using the Brain Health Index (BHI), a comprehensive composite battery of 23 distinct behavioral and cognitive assessments. In a study published in *Nature Scientific Reports* evaluating 4,000 individuals ranging in age from 18 to 93, participants who completed targeted cognitive training strategies achieved an average 7.9% improvement across cognitive assessments alongside measurable, deeper activation within the prefrontal cortex.

- **Institution Focus:** Translational research targeting lifelong cognitive performance, deficits, and interventional protocols. - **The BrainHealth Project:** 10-year longitudinal study across all adult age brackets at 6-month intervals, aiming for 100,000 participants. - **Assessment Metric:** Brain Health Index (BHI), derived from 23 behavioral and cognitive assessments. - **Clinical Validation:** Published trial in *Nature Scientific Reports* demonstrating an average 7.9% cognitive gain and increased prefrontal cortex activation across 4,000 participants (ages 18–93).

---

[02:51](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=171s) Executive Function, Neuroplasticity, and the Core Question of Technology

At the center of human agency lies executive function, seated primarily within the prefrontal cortex. This region governs planning, abstract reasoning, critical thinking, self-direction, and creative synthesis. When individuals engage in structured cognitive training targeting these executive faculties, physical and functional changes occur across neural pathways—a foundational dynamic known as neuroplasticity. The operating principle is straightforward: neural connections that are consistently challenged and exercised are reinforced and strengthened, whereas unused neural pathways progressively atrophy.

When addressing whether emerging technology is fundamentally beneficial or harmful to brain health, Dr. Sandra Chapman consistently responds: "Yes, it's both." Whether technology bolsters or degrades neural capacity depends entirely on the behavioral patterns of adoption. While AI can rapidly level the playing field by closing the skills gap between junior and senior professionals, habitual overreliance risks bypassing the very executive processes needed to sustain independent analytical capabilities.

```mermaid flowchart TD accTitle: Cognitive Amplification Versus Cognitive Delegation Pathways subgraph HumanInput["Executive Agency"] A["Core Problem / Task"] --> B{"Adoption Approach"} end

subgraph Amplification["Cognitive Amplification"] B -->|"Originate First"| C["Deep Thought Work<br/>(Drafts, Reasoning, Framing)"] C --> D["Targeted AI Acceleration<br/>(Editing, Details, Tedium)"] D --> E["Strengthened Prefrontal Networks"] end

subgraph Delegation["Cognitive Delegation (Debt Cycle)"] B -->|"Offload Early"| F["Outsource Primary Thinking<br/>(Instant Summaries, Outlines)"] F --> G["Cognitive Atrophy & Homogenization"] G --> H["Deepening AI Dependency"] H -.->|"Feedback Loop"| F end ```

---

[04:37](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=277s) The Cognitive Debt Cycle and Amplification vs. Delegation

When individuals routinely delegate initial cognitive heavy lifting to artificial intelligence, they enter what researchers term the "cognitive debt cycle." By circumventing personal cognitive effort, users experience a progressive decline in independent critical reasoning and problem-solving. This phenomenon was highlighted in an essay-writing study conducted at the MIT Media Lab: participants heavily dependent on AI generated structurally sound essays with high initial marks for surface creativity, yet their work suffered from severe conceptual homogenization. Furthermore, 83% of the AI-reliant group were entirely unable to quote directly from or provide an outline of their own submitted essays.

Because avoiding AI entirely is not a viable strategy, users face a fundamental fork in the road:

1. **Cognitive Amplification:** The user takes ownership of the primary thought work—authoring the rough draft, formulating the initial synthesis, or reading source material directly—before employing AI to handle formatting, surface editing, administrative tedium, or granular project detail. 2. **Cognitive Delegation:** The user offloads higher-order synthesis and conceptual formulation directly to the model, gradually losing the capacity for sustained executive thinking.

As advanced foundational models become ubiquitous within every standard operating system, competitive advantage will no longer stem from access to the tool itself. The true differentiator will be the strategic discipline to maintain personal executive effort, using AI strictly to accelerate execution while keeping core reasoning inside human biological networks.

---

[08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) A Four-Part Framework for Brain-Healthy AI Adoption

To prevent cognitive decline and preserve personal agency while leveraging AI, a structured four-part behavioral framework is proposed:

1. **Basic Neuroscience Literacy:** Paralleling nutritional and physical fitness literacy, understanding how mental habits impact neural structures allows individuals to make conscious, protective choices about when to exert effort and when to automate. 2. **Originate Before You Delegate:** Always perform the foundational framing, drafting, and problem formulation prior to prompting an AI. Preserving this initial friction ensures that the prefrontal cortex receives the metabolic workout necessary to maintain cognitive health. 3. **Disengage Purposefully:** Grounded in advice from design and art direction professor Lee Whitmarsh to continuously "fill your visual well," this principle emphasizes digesting original materials directly—such as reading a complete book or studying cultural trends firsthand—rather than consuming pre-digested summaries. Direct exposure builds rich internal associative networks that external summaries cannot supply. 4. **Reflect Wisely:** Establish an end-of-day reflective practice by auditing the completed task list and honestly evaluating: *"Who did more of the effort there, me or AI?"* If AI consistently shoulders the majority of deep cognitive labor day after day, users must deliberately throttle back their AI reliance to rebuild their neural cognitive reserves against future neurological stress.

---

[11:41](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=701s) Expanding Research: Social Dependencies, SMART Training, and Brain-Healthy AI Agents

The Center for BrainHealth is actively updating its research infrastructure to address the cognitive and behavioral shifts driven by artificial intelligence. Within its ongoing ten-year longitudinal study, the team is embedding dedicated AI-tracking telemetry to evaluate how the frequency, type, and sophistication of AI interactions correlate with long-term Brain Health Index trajectories.

```mermaid flowchart LR accTitle: Center for BrainHealth AI Research Integration subgraph Tracking["Telemetry & Monitoring"] A["AI Tracking Metrics<br/>(Frequency, Type, Depth)"] B["Charisma Platform<br/>(Social Cognition & ISERT)"] end

subgraph CoreEngine["Longitudinal Science"] C["Brain Health Index (BHI)<br/>10-Year Study Cohort"] end

subgraph Interventions["Applied Protocols"] D["SMART Cognitive Training<br/>+ AI Best Practices"] E["Precision Brain Health AI Agents"] end

A --> C B --> C C --> D C --> E ```

Beyond isolated cognitive tasks, the center is investigating the social ramifications of AI use. As individuals experience increasing social isolation, dependency on artificial agents for companionship and interaction is accelerating. Using the center's *Charisma* platform and its proprietary social cognition assessment metric, the team is evaluating how varying patterns of AI engagement influence real-world interpersonal competence.

Finally, the center is updating its evidence-based cognitive training protocol—Strategic Memory Advanced Reasoning Training (SMART)—to include best practices for brain-healthy AI adoption. Just as widespread physical automation historically necessitated the invention of athletic gyms to ensure deliberate physical exertion, cognitive automation now requires formal cognitive conditioning. The center is currently applying these insights to build precision brain health models and next-generation AI agents that natively encourage users to follow brain-healthy workflows.

[13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) AI Guardrails Against Cognitive Offloading and the Economics of Brain Health

To prevent passive over-reliance on artificial intelligence, agentic workflows can be engineered with built-in "guardrails for your own cognitive offloading." Rather than taking over cognitive tasks entirely, an AI system can be structured to prompt the user to perform the primary mental heavy lifting first—such as demanding, "you give me the first draft and then I'll do the improvement." By intentionally embedding these interaction patterns, developers and organizations can force users to continue exercising their own cognitive work rather than outsourcing it completely.

Driving widespread adoption of these cognitive practices ultimately requires demonstrating tangible return on investment, because "money talks." Dr. Andrew Nevin, an economist collaborating with Dr. Jackie Gamino at the Center for BrainHealth, previously demonstrated that training middle school students in critical thinking skills produces measurable increases in their lifetime earning potential. Building on that foundation, Dr. Nevin is now examining how proactive brain health interventions generate broader workforce, societal, economic, and fiscal benefits for organizations that choose to adopt them.

```mermaid flowchart TD accTitle: Cognitive Guardrails and Economic Value Flow subgraph HumanAI ["Intentional AI Design"] A["User Input / Initial Drafting"] --> B["Cognitive Guardrail Prompts"] B --> C["AI Refinement & Improvement"] end subgraph Outcomes ["Long-Term Value"] C --> D["Critical Thinking & Brain Health Retention"] D --> E["Workforce Productivity & Higher Lifetime Earnings"] D --> F["Measurable Societal & Fiscal Returns"] end ```

[15:06](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=906s) Translational Research and Organizational Partnerships

As a not-for-profit translational research institute embedded within The University of Texas at Dallas, the Center for BrainHealth cannot scale these solutions commercially on its own. While the center has validated measurement protocols and proven cognitive interventions that demonstrate clear individual and economic improvements, bringing these tools to market requires active collaboration with outside industry partners.

- **Key Research Collaborators:** Dr. Andrew Nevin (economist) and Dr. Jackie Gamino (Center for BrainHealth). - **Institutional Role:** The Center for BrainHealth functions as a university-based translational research institute developing evidence-backed measures and cognitive interventions. - **Resources & Study Access:** Additional research information, study participation, and organizational inquiries can be accessed through `centerforbrainhealth.org` and `brainhealthpro.org`.

## Notable quotes > AI is the most powerful cognitive tool ever built. The question is whether you are using it to strengthen your brain or quietly letting it do the thinking for you. — Aaron Tate

> How we adopt it may matter more than whether we adopt it at all. — Aaron Tate

> [02:51](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=171s) Yes, it's both. — Dr. Sandra Chapman

> [04:37](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=277s) cognitive debt cycle — Aaron Tate

> [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) Originate Before You Delegate — Aaron Tate

> [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) fill your visual well — Lee Whitmarsh

> [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) Who did more of the effort there, me or AI? — Aaron Tate

> [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) guardrails for your own cognitive offloading — Aaron Tate

> [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) you give me the first draft and then I'll do the improvement. — Aaron Tate

> [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) money talks. — Aaron Tate

> what is effortless, automated thinking doing to the cognitive infrastructure everyone is running on but nobody is maintaining? — Aaron Tate

## Mentioned Aaron Tate, Ai4, Brain Health Index (BHI), brainhealthpro.org, Center for BrainHealth, centerforbrainhealth.org, Charisma, Dr. Andrew Nevin, Dr. Jackie Gamino, Dr. Sandra Chapman, Lee Whitmarsh, MIT Media Lab, Nature Scientific Reports, Strategic Memory Advanced Reasoning Training (SMART), The BrainHealth Project, The University of Texas at Dallas

## Video description Aaron Tate, Director, Center for BrainHealth, The University of Texas at Dallas

AI is the most powerful cognitive tool ever built. The question is whether you are using it to strengthen your brain or quietly letting it do the thinking for you.

Research is documenting measurable neurological changes from AI use, including reduced cognitive control, eroding critical thinking, and risks to social cognition that are only beginning to be studied. But the evidence is equally clear that when AI is used intentionally, it amplifies performance. How we adopt it may matter more than whether we adopt it at all.

At the Center for BrainHealth at UT Dallas, we have spent 30 years studying how the brain performs, how it declines, and how it recovers. We built validated tools to measure cognitive performance holistically, and we are now applying that same rigor to track what AI use is actually doing to brain health over time. This is not a new problem for us.

Attendees will leave with a clear model for understanding how AI use patterns drive divergent cognitive outcomes and four evidence-based practices to protect and strengthen their cognitive edge.

The entire tech world is focused on adopting AI faster. This talk asks the question no one is asking: what is effortless, automated thinking doing to the cognitive infrastructure everyone is running on but nobody is maintaining?

[Source](https://www.youtube.com/watch?v=-Rpu7N_nK74)

中文简报

专题讨论 6 - 4：AI 时代的大脑健康：重拾自主权 Aaron Tate 解释了将认知被动委托给 AI 如何引发有害的“认知负债周期”，导致前额叶皮层萎缩、概念同质化以及批判性思维受损。结合 The BrainHealth Project 的研究与纵向追踪数据，他区分了被动委托与有意的认知增强——在后者中，用户在利用 AI 执行任务之前仍保留核心的概念推理。Tate 概述了一个由四部分组成的行为框架：神经科学素养、先自主构思再委托、有意识的抽离以及反思性审查，并提出了自主性护栏和更新版 SMART 认知方案，以建立可持续的神经储备。 德克萨斯大学达拉斯分校脑健康中心通过大脑健康指数（BHI）追踪长期认知表现，其发表在《自然-科学报告》上的临床研究证实，针对性认知训练可使认知评估表现平均提升 7.9%，并加深前额叶皮层的激活。 (0:04) 频繁将核心思考外包给 AI 会陷入“认知债务周期”并导致批判性思维退化；MIT 媒体实验室的一项论文研究显示，高度依赖 AI 的参与者中有 83% 甚至无法引用或复述自己提交论文的大纲。 (4:37) 随着大模型普及，竞争优势将不再来自工具本身，而取决于保持自主执行思考的自律：应当坚持“认知放大”（先自主完成深度框架再由 AI 辅助落地），而非“认知委托”（直接外包核心合成与判断）。 (4:37) 保护大脑健康的四项 AI 使用法则包括：建立基础神经科学常识、先自主构思再交付 AI、脱离预加工摘要直接研读原始资料，以及每日反思自己在任务中实际投入的智力比例。 (8:31) 如同工业时代的体力自动化催生了体育健身房，认知自动化也需要结构化的认知训练与防过度外包的交互护栏，促使用户在 AI 介入优化前必须先锻炼自身的思考肌肉。 (13:58)

Reading summary

While artificial intelligence accelerates output and can narrow professional skills gaps, researchers at the Center for BrainHealth emphasize that routine cognitive delegation risks neural atrophy and weakened executive function. Sustaining human agency requires adopting intentional amplification practices that keep core reasoning within biological neural networks before delegating tasks to automated systems. Prefrontal Plasticity and Cognitive Health Measurement (0:04) Executive function—governing abstract reasoning, planning, critical thinking, and synthesis—resides primarily in the prefrontal cortex and operates through neuroplasticity. Because neural pathways strengthen when actively exercised and atrophy when neglected, Center for BrainHealth founder Dr. Sandra Chapman argues that emerging technology is simultaneously beneficial and harmful: its ultimate impact depends entirely on behavioral adoption patterns. To quantify these dynamics, the center established The BrainHealth Project, a 10-year longitudinal study evaluating a target cohort of 100,000 healthy adults at six-month intervals using the Brain Health Index (BHI), a composite battery of 23 behavioral and cognitive assessments. In a trial published in Nature Scientific Reports covering 4,000 participants aged 18 to 93, individuals who practiced targeted cognitive training protocols demonstrated an average 7.9% improvement across cognitive measures, alongside measurable increases in prefrontal cortex activation. Cognitive Debt and the Delegation Dilemma (4:37) Routinely outsourcing initial intellectual effort places users into what researchers call a cognitive debt cycle, leading to the progressive erosion of independent reasoning. An essay-writing study conducted at the MIT Media Lab demonstrated this risk: while AI-reliant participants submitted structurally sound essays that scored high in surface creativity, their output suffered from marked conceptual homogenization. Moreover, 83% of the AI-reliant cohort were incapable of outlining or directly quoting from their own submitted work. To avoid this cycle as foundational AI models integrate across operating systems, workers face a structural choice between cognitive delegation and cognitive amplification. Delegation offloads conceptual formulation and higher-order synthesis directly to models, degrading analytical capacity over time. Conversely, amplification demands that humans retain ownership of core framing, rough drafting, and primary synthesis, deploying AI strictly to assist with formatting, administrative friction, and execution speed. A Four-Part Behavioral Framework for AI Adoption (8:31) To counter cognitive degradation while utilizing AI tools, researchers outline a four-part behavioral protocol. First, individuals require foundational neuroscience literacy to understand how automated habits physically reshape brain structures. Second, users should follow the rule to originate before delegating, preserving the initial cognitive friction of drafting and problem framing to ensure the prefrontal cortex receives necessary metabolic engagement. Third, users must purposefully disengage by absorbing unsummarized source materials directly—such as reading complete books or analyzing primary trends—reflecting design professor Lee Whitmarsh's principle of continuously filling one's internal visual and associative well. Fourth, workers should adopt a daily reflective audit, reviewing completed tasks to assess whether human thought or automated software performed the majority of cognitive heavy lifting, adjusting reliance whenever deep personal effort falls behind. Cognitive Guardrails, Telemetry, and Economic Valuation (11:41) The Center for BrainHealth is updating its longitudinal telemetry to record the frequency, depth, and type of AI interactions against long-term BHI metrics, while using its Charisma platform to evaluate how AI companion interactions affect interpersonal social cognition. Concurrently, the center is revising its Strategic Memory Advanced Reasoning Training (SMART) framework to model precision AI agents with built-in cognitive guardrails. Rather than automating full assignments, these workflows mandate that users supply initial drafts or arguments before the AI delivers refinements. Broader workplace adoption depends on demonstrating economic returns. Economist Dr. Andrew Nevin, collaborating with Dr. Jackie Gamino, previously established that critical thinking interventions in middle school students yield measurable gains in lifetime earnings. Nevin is expanding this research to assess the organizational, fiscal, and workforce productivity dividends realized when enterprises proactively implement brain health protocols. Because the Center for BrainHealth operates as a university-based non-profit translational research institute, scaling these validated cognitive training protocols and AI guardrail architectures will require formal partnerships with commercial industry, while the broader societal impacts of AI dependency on real-world social cognition remain under active study.

精炼阅读摘要

德克萨斯大学达拉斯分校脑健康中心的研究指出，人工智能对人脑健康的影响取决于具体的使用模式：过度依赖会导致执行功能萎缩并陷入“认知债务”，而主动承担核心推理并善用技术作为放大器则有助于维持神经可塑性。 脑健康评估与神经可塑性机制 (0:04) 德克萨斯大学达拉斯分校脑健康中心（Center for BrainHealth）由 Sandra Chapman 博士创立，三十余年来专注于全年龄段健康人群的认知潜力与脑功能保护研究。该机构的核心科研项目“脑健康计划”是一项为期十年的纵向研究，旨在追踪10万名健康成年受试者，每隔6个月利用涵盖23项行为与认知评估的“脑健康指数”（BHI）进行多维测定。发表在《自然·科学报告》（Nature Scientific Reports）上的一项涉及4,000名18至93岁受试者的试验显示，接受针对性认知策略训练的个体在认知测试中平均取得了7.9%的提升，并伴随前额叶皮层更深层次的激活。 Chapman 博士强调，前额叶皮层负责主导规划、抽象推理、批判性思维与自主方向等核心执行功能，其运作严格遵循神经可塑性法则——经常调动和挑战的神经通路会不断强化，而缺乏刺激的通路则会逐渐退化。面对技术利弊的争论，Chapman 认为AI既能迅速缩小初级与资深人员之间的专业差距，也可能因习惯性滥用而使用户绕过维持独立分析能力所必需的执行功能，其最终结果完全取决于人类采纳该技术时的行为模式。 认知债务陷阱与“放大对代劳”抉择 (4:37) 当使用者习惯性地将初阶高负荷思考直接移交给AI时，就会陷入研究人员所谓的“认知债务循环”。麻省理工学院媒体实验室（MIT Media Lab）的一项论文写作实验佐证了这种负面影响：高度依赖AI完成写作的受试组虽然产出了结构完整且在初期表层创造力评分较高的文章，但其内容表现出严重的同质化倾向；更关键的是，高达83%的AI依赖组参与者在完成提交后，完全无法准确引用或口头列出自己所提交论文的大纲，显示出其核心推理并未经过自身大脑的加工。 鉴于彻底规避AI并不具备现实可行性，研究人员将当下的应用策略划分为两个方向：“认知放大”（Cognitive Amplification）与“认知代劳”（Cognitive Delegation）。在放大模式中，用户始终主导初步框架构思、原始逻辑综合与第一版草稿撰写，仅让AI负责排版格式化、文字润色或琐碎细节；而在代劳模式中，用户把高阶综合与概念成型全盘外包。随着基础大模型逐渐整合进通用操作系统，未来的竞争优势将不再取决于是否掌握工具，而取决于能否自律地维持高强度的生物大脑神经运算。 维持大脑健康的四项行为框架 (8:31) 为抵御认知退化并保持个人自主性，研究提出了一套包含四个维度的日常行为框架：第一，建立基础神经科学素养。如同理解营养摄入与体能训练一样，认知到心智习惯对大脑物理结构的重塑作用，有助于用户清醒决定何时调动脑力、何时交由自动化。第二，“先原创，后委托”（Originate Before You Delegate）。在向AI输入提示词之前，必须亲自主导问题的定义、框架梳理与初步构思，保留必要的认知阻力以维持前额叶皮层的代谢训练。 第三，有目的地抽离（Disengage Purposefully）。该原则借鉴了设计学教授 Lee Whitmarsh “持续注满视觉之井”的观点，强调个体必须直接精读一手书籍原著或实地观察文化动态，而非长期依赖经过AI压缩提炼的摘要，以构建外部简报无法提供的大脑内部联想网络。第四，明智反思（Reflect Wisely）。建议用户在一天结束时复盘任务清单，审视核心智力劳动主要是由自己完成还是由AI承担；若发现AI连续占据主导，则需主动收缩使用范围，以重建对抗神经压力的认知储备。 认知护栏设计与长期经济效益 (11:41) 脑健康中心正在对其长期研究架构进行升级，将AI交互遥测指标整合进持续十年的跟踪队列中，以观察AI使用频率、交互类型及复杂度对BHI指标的长期影响，并通过其 Charisma 平台与社交认知量表评估AI虚拟陪伴对人际交往能力的潜在改变。针对认知卸载风险，研究人员主张在智能体工作流中构建技术层面的“防认知卸载护栏”（Guardrails Against Cognitive Offloading），例如在软件底层设置交互规则，强制要求用户必须先提供原创初稿，系统随后才介入优化，从而迫使用户保持认知锻炼。 推动全社会采纳这些健康脑力机制需要验证明确的投资回报。经济学家 Andrew Nevin 与脑健康中心 Jackie Gamino 博士前期的合作研究证实，对初中生开展批判性思维训练能显著提高其终身收入潜力。在此基础上，Nevin 博士正在进一步量化在组织与劳动力层面实施积极脑健康干预所带来的生产力提升、组织留存及宏观财政效益。 脑健康中心作为大学下属的非营利性转化研究机构，已建立起经过临床验证的测量体系与干预手段，但其自身无法直接推动这些方案的大规模商业落地，必须依赖与外部工业界的合作。在追求即时效率最大化的技术浪潮中，如何说服企业和组织主动部署可能“增加认知摩擦”的干预护栏，仍是实现技术与人脑平衡发展的关键考验。

Detailed notes — zh

## Background 该视频由 YouTube 频道“Ai4”于 2026 年 8 月 18 日发布，主讲人为德克萨斯大学达拉斯分校脑健康中心（Center for BrainHealth）主任 Aaron Tate。内容简介指出，AI 作为强大的认知工具正对大脑产生可测量的神经学改变，演讲旨在剖析不同 AI 使用模式对脑健康的影响，并提供四项基于循证医学的护脑实践策略以重塑人类的认知自主权。 ## TL;DR 演讲依托脑健康中心 30 年的研究积淀与多项纵向实证数据，指出大脑遵循“用进废退”的神经可塑性铁律，过度依赖 AI 会导致高阶思维退化并陷入“认知债务循环”。发言人重点辨析了主动掌控核心推导的“认知放大”与全盘外包思维的“认知委托”两种路径，并提出“神经科学素养、先构思后外包、目的性抽离、日终复盘”四项防护准则。同时，研究团队正探索在 AI 智能体中内置强制性认知护栏，并结合经济学研究揭示了批判性思维训练对个人终身收益及组织效能的实质回报。 ## Atmosphere & style 整场演讲呈现出严谨扎实、发人深省且兼具警示意味的学术对话基调。演讲者摒弃了极端的唯技术论或盲目悲观情绪，以翔实的神经影像学研究和前额叶激活数据为基石，结合个人设计学习经历与生动的日常比喻娓娓道来。叙事节奏层层递进，既有对人类认知被无声空心化的冷峻洞察，又给出了富有建设性的防护方案，营造出一种促使听众深刻反思自身思维主动权的理性氛围。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=4s) 德克萨斯大学达拉斯分校脑健康中心开展的“脑健康计划”依托长达10年的纵向追踪，证实有针对性的认知训练可使认知表现平均提升7.9%，并显著激活前额叶皮层。 - [00:04](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=4s) 前额叶皮层所承载的执行功能是人类自主性的核心，大脑神经可塑性遵循严酷的“用进废退”法则：得到锻炼的回路被强化，被技术替代而长期闲置的连接则走向退化。 - [03:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=238s) AI 对脑健康的影响取决于使用方式，过度依赖 AI 会让人陷入“认知债务循环”——主动动脑越少导致脑力越退化，进而对技术产生更深的依赖。 - [03:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=238s) 麻省理工学院媒体实验室研究证实，重度依赖 AI 撰写文章会导致内容概念高度同质化，高达 83% 的参与者甚至无法复述出自己生成文章的核心大纲或原句。 - [06:42](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=402s) 面对 AI 的核心分歧在于“认知放大”还是“认知委托”：前者由人主导核心推导与初稿、仅让 AI 负责辅助润色；后者将思考全权外包，使高阶思维能力与自主性逐渐空心化。 - [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) 维护脑健康需要具备基础神经科学素养，理解用进废退与神经可塑性机制，像管理饮食热量与营养一样建立克制使用 AI 的认知自律。 - [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) 实践“先构思再委托”原则，在向 AI 求助前务必逼迫自己独立完成原始洞察与初版框架，确保前额叶皮层获得必不可少的阻力锻炼。 - [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) 通过有目的地主动断网抽离去啃读一手难啃资料以充实认知储备，并在每日工作结束前坦诚复盘人机脑力支出占比，动态纠偏对 AI 的依赖。 - [11:41](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=701s) 脑健康中心在纵向科研中增设了 AI 交互频率与提示词成熟度等新维度，并借助 Charisma 测评体系深入评估用 AI 填补人际空虚对真实社交认知网络的潜在冲击。 - [11:41](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=701s) 如同工业机械化免除重体力劳动后催生了身体健身房，中心正结合 SMART 认知训练体系构建精准脑健康模型，将脑保护策略直接内置于软件交互中。 - [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) 人机交互设计必须设立防止过度“认知卸载”的护栏，引导或强制用户先产出思考初稿再由 AI 介入优化，避免自动化输出导致用户大脑深度思考功能退化。 - [14:15](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=855s) 实证研究显示青少年期接受批判性思维干预能显著提升其一生潜在收入，将科学的脑健康训练扩展至企业和组织也将带来切实的生产力与宏观经济回报。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=4s) 脑健康研究的基石：神经可塑性与认知自主权 人工智能（AI）无疑是当今提升生产力、工作效率甚至激发创造力的强大工具，但我们必须停下脚步，审视它对人类脑健康可能带来的负面冲击。德克萨斯大学达拉斯分校脑健康中心（Center for BrainHealth）由桑德拉·查普曼博士（Dr. Sandra Chapman）于 30 多年前创立，是一家专注于各年龄段大脑健康、功能表现与认知缺陷转化的转化医学研究机构。该中心的旗舰项目“脑健康计划”（The BrainHealth Project）是一项为期 10 年的纵向研究，旨在招募 10 万名健康成年人受试者，每隔 6 个月进行一次追踪评估，其中一部分人群还会接受直接的大脑神经影像学检查。为了系统量化这一进程，研究团队构建了“脑健康指数”（Brain Health Index），涵盖 23 项独立的认知与行为评估。近期发表在《自然·科学报告》（Nature Scientific Reports）上的一项研究显示，在 4000 名年龄跨度为 18 岁至 93 岁的参与者中，接受认知训练策略后，其认知评估成绩平均提升了 7.9%，并且在前额叶皮层观察到了更深层的神经激活。 前额叶皮层承载的执行功能——包括规划、推理、批判性思维与创造力——是人类自主性（Agency）的核心，它赋予了个体自我导向的行动力。脑健康中心的研究表明，当参与者接受针对执行功能的定向训练时，大脑会发生可测量的积极改变，这正是“神经可塑性”（Neuroplasticity）的直观体现。神经可塑性的基本法则是极其严酷的：频繁锻炼的神经连接会得到强化，而长期闲置的神经连接则会逐渐萎缩。 - **脑健康计划规模**：10 年纵向追踪，每 6 个月评估一次，目标招募 100,000 名跨年龄段健康成年人，部分配合神经影像检测。 - **《自然·科学报告》实证成果**：覆盖 4,000 名受试者（18-93 岁），认知评估平均提高 7.9%，前额叶激活显著增强。 - **神经可塑性铁律**：用进废退——得到锻炼的突触回路得以巩固，被替代而废弃的连接则走向退化。 --- ### [03:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=238s) 认知债务循环：当 AI 替代深度思考 面对技术浪潮，人们常常向查普曼博士提出一个经典问题（过去针对智能手机，如今聚焦于 AI）：“技术对脑健康究竟是有益还是有害？”她的回答始终如一：“两者皆是（Yes, it's both）。”技术的利弊完全取决于我们如何使用它。虽然 AI 能够显著抹平初级从业者与资深专家之间的技能鸿沟、大幅提高工作产出，但如果长期过度依赖 AI，个体自身的逻辑推理与批判性思考能力就会遭到不可逆的削弱，进而陷入所谓的“认知债务循环”（Cognitive Debt Cycle）。 这种债务循环表现为一种恶性递进：个体越频繁使用 AI，对其依赖程度就越高；越是依赖，自己主动动用认知功能的频率就越低，从而导致自身脑力水平明显退化。麻省理工学院媒体实验室（MIT Media Lab）的一项实验清晰地揭示了这一代价：在该研究中，重度依赖 AI 撰写文章的群体虽然在语法结构和部分创意指标上表现优异，但内容概念却出现了极其严重的同质化趋势；更令人震惊的是，这些高度依赖 AI 的参与者中，竟有高达 83% 的人甚至无法复述出自己文章中的原句或列出核心大纲。这表明他们虽然交出了一篇漂亮的作品，但大脑并未真正参与深度思考。 ```mermaid flowchart TD accTitle: 认知债务循环运作机制 subgraph S1["依赖的起点"] A["高频过度使用 AI"] --> B["思维主动权让渡给机器"] end subgraph S2["功能退化阶段"] B --> C["关键脑区闲置<br/>（缺少前额叶认知锻炼）"] C --> D["自主推理与批判性思维退化"] end subgraph S3["恶性固化"] D --> E["对 AI 依赖度进一步加剧"] E --> A end ``` --- ### [06:42](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=402s) 认知放大与认知委托的二元抉择 面对 AI 时代，完全拒绝采用 AI 显然不是可行的选项，真正关键的抉择在于使用模式的根本分歧：你是在进行“认知放大”（Cognitive Amplification），还是在进行“认知委托”（Cognitive Delegation）？在认知放大的模式中，用户始终将硬核的思维工作握在自己手中——自己写第一版草稿、独立完成初步分析、亲自通读消化书籍而不是直接索要摘要；随后，才调动 AI 去处理琐碎繁重的辅助工作、文本润色或在既定框架下补足细节。相反，认知委托则直接滑向了认知债务循环，让人逐步放弃独立思考，长此以往便丧失了从事高阶思维的能力。 当顶尖的大语言模型最终无缝集成到每个人的操作系统底座时，未来的核心竞争力不再是谁拥有更强的工具，而是谁能以战略性、主动性的方式使用 AI。那些能够通过认知放大将 AI 视作催化剂、自身承担更高强度思维训练的人，才能够实现认知跃迁；而将思维外包给机器的人，其认知能力则在无声无息中被空心化。 | 模式对比 | 认知放大（Cognitive Amplification） | 认知委托（Cognitive Delegation） | |:--- |:--- |:--- | | **思维主体** | 人类自身主导核心推导、立意与初步构思 | 将起始概念构思与思考全权外包给 AI | | **工作分配** | 亲自完成初稿与原始洞察，由 AI 辅助润色、排查琐碎细节 | 直接让 AI 生成成品，个人仅做表面批改或被动接受 | | **神经影响** | 强化前额叶皮层连接，锻炼并保护认知储备 | 陷入认知债务循环，高阶思维能力与自主权萎缩 | | **典型案例** | 研读整本书形成认知后，再用 AI 对比要点 | 从未阅读原著，仅依靠 AI 一键生成的摘要应付差事 | --- ### [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) 维护脑健康的四项 AI 使用守则 为了让大众在享受技术红利的同时保护大脑健康，脑健康中心提出了四条基于循证神经科学的使用准则。第一是“基础神经科学素养”（Basic Neuroscience Literacy）。正如掌握卡路里与营养学知识能帮助我们在面对高糖甜点时做出克制决策一样，理解大脑的神经可塑性机制，能促使我们在工作中自觉警惕过度外包思考的诱惑。第二是“先构思再委托”（Originate before you delegate / Originate before you offload）。在向 AI 求助前，务必逼迫自己完成最初的草稿和核心框架，确保前额叶皮层获得必不可少的阻力训练。 第三是“有目的地主动脱离”（Disengage purposefully）。演讲者回顾了自己大学本科攻读设计与艺术指导时，教授李·惠特马什（Lee Whitmarsh）给出的建议：设计师必须主动“填满自己的视觉之井”（fill our visual well），去广泛涉猎电影、书籍、音乐与流行文化的时代风貌，唯有脑中具备丰沛的原始养分，才能在客户提出模糊需求时准确领会并做出独创反应。同样，主动关掉 AI、亲自去啃难啃的书籍，正是为了在脑海中沉淀真正的神经网络资产。第四是“明智地反思检视”（Reflect wisely）。演讲者分享了自己的日常习惯：每天下班前审视当日完成的任务清单，坦诚自问：“今天的工作究竟是我投入的心力多，还是 AI 贡献得更多？”如果连续多天发现天平完全倒向 AI，就必须紧急拉起警报、收紧对 AI 的依赖，以此维护面对突发神经退行或健康意外时至关重要的“认知储备”（Cognitive Reserve）。 - **四项循证实践**： 1. **神经科学素养**：洞悉用进废退机制，建立像管理热量摄入一样的认知自律。 2. **先构思后外包**：保留初稿撰写与破题权，避免将核心脑力活动直接让渡。 3. **目的性抽离**：有计划地断开技术辅助，自主消化第一手资料以扩充思想源泉。 4. **日终复盘反省**：每日复盘人机脑力支出占比，动态校准，坚决捍卫长期神经认知韧性。 --- ### [11:41](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=701s) 纵向科研前沿：多维度量化人机交互与精准干预 脑健康中心并未止步于理论倡导，而是将 AI 时代的新变量直接纳入了其庞大的科研体系中。在现有的 10 年期纵向研究中，研究团队正在增设专门针对 AI 使用行为的追踪指标，涵盖使用者与 AI 交互的频率、具体应用场景以及提示词与思维交互的成熟度（Sophistication），并将其与受试者的综合脑健康指数展开横向与纵向比对。此外，针对当下日益凸显的“社交依赖”与社交隔离现象，该中心正借助其自主研发的社交认知测评平台 Charisma 及其下属的“Isert”社交能力评估体系，深入测算人类用 AI 填补人际空虚对实际社交认知网络带来的潜在冲击。 针对这套现实挑战，脑健康中心进一步升级了原有的认知训练体系。正如人类在工农业机械化彻底免除重体力劳动后、为了维持身体机能而不得不发明“健身房”一样，该中心正将其著名的“SMART 认知训练体系”（Strategic Memory Advanced Reasoning Training）与脑健康 AI 最佳实践融为一体。与此同时，团队正在开发针对个体的“精准脑健康模型”与健康型 AI 智能体，探索如何将“先构思后委托”等脑健康防线直接编码到软件交互逻辑中，从源头上协助人类重新夺回属于自己的思维控制权。 ### [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) 构建认知防护栏：让 AI 促发自主思考而非过度认知卸载 在人机交互与 AI 智能体的设计上，必须建立起防止人类过度进行“认知卸载”（Cognitive Offloading）的护栏机制。理想的 AI 智能体不应完全代劳所有的思考过程，而是应当倒逼用户先进行自主的认知活动。 智能体应当在交互中向用户发出引导性提示，例如明确设定规则：“请你先提供初稿，随后我再来进行优化与润色”（"you give me the first draft and then I'll do the improvement."）。通过在底层模型与智能体中内置这种强制性机制，能够促使用户完成属于自己的大脑认知训练，避免大脑在毫不费力的自动化输出中逐渐退化。 ```mermaid flowchart TD accTitle: AI Cognitive Guardrail Flow subgraph S1["传统被动依赖模式"] A["用户提出需求"] --> B["AI 全包全办完成初稿"] B --> C["人类认知退化与被动接收"] end subgraph S2["具备认知防护栏的协作模式"] D["用户提出需求"] --> E["AI 引导：要求用户提供思考初稿"] E --> F["用户完成核心认知加工"] F --> G["AI 介入进行迭代与精炼"] end ``` --- ### [14:15](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=855s) 脑健康的经济价值：从终身收入提升到组织级生产力回报 要让重视脑健康与认知训练的理念获得更广泛的市场采纳，关键在于证明其明确的经济价值——正如俗话所说，“金钱最具说服力”（"money talks"）。达拉斯德克萨斯大学脑健康中心（Center for BrainHealth）正通过严谨的研究将神经科学与经济效益紧密挂钩。 经济学家安德鲁·内文博士（Dr. Andrew Nevin）与脑健康中心的杰基·加米诺博士（Dr. Jackie Gamino）此前开展了深入合作研究，结果表明：在中学生阶段进行针对性的批判性思维技能训练，能够显著且可量化地提高他们一生的潜在收入水平。如今，内文博士正在将研究维度拓展至更广泛的领域，深入探讨企业和组织采纳脑健康训练实践后，如何在劳动力素质、社会效益、宏观经济以及财政收入等层面带来实打实的积极回报。 作为一家依托于大学的非营利性转化医学与科学研究机构，脑健康中心本身无法直接将技术大规模商业化，因此极为需要具备落地能力的组织合作伙伴共同将研究成果推向市场。中心目前已经具备完善的评估量表与干预训练方案，并充分证实了这些方案不仅能实现个人认知层面的提升，还能为组织带来切实的经济与财政效益。 - **关键研究成果与合作学者**：经济学家安德鲁·内文博士（Dr. Andrew Nevin）携手杰基·加米诺博士（Dr. Jackie Gamino），证实了青少年批判性思维干预对未来终身收入的显著提振作用。 - **转化研究与商业合作诉求**：脑健康中心呼吁产业界组织展开深度合作，共同将科学验证的脑健康评估指标与认知干预方案规模化落地。 - **获取更多研究资讯**：可访问官方网站 `centerforbrainhealth.org` 或参与其研究项目的专属平台 `brainhealthpro.org` 获取更多信息并建立联络。 ## Notable quotes > [03:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=238s) Yes, it's both. — Dr. Sandra Chapman > [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) Originate before you delegate — Aaron Tate > [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) Originate before you offload — Aaron Tate > [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) fill our visual well — Lee Whitmarsh > [08:31](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=511s) Reflect wisely — Aaron Tate > [13:58](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=838s) you give me the first draft and then I'll do the improvement. — Aaron Tate > [14:15](https://www.youtube.com/watch?v=-Rpu7N_nK74&t=855s) money talks — Aaron Tate ## Mentioned Aaron Tate, Center for BrainHealth, The University of Texas at Dallas, Ai4, Dr. Sandra Chapman, The BrainHealth Project, Nature Scientific Reports, MIT Media Lab, Lee Whitmarsh, Charisma, Isert, SMART 认知训练体系, Dr. Andrew Nevin, Dr. Jackie Gamino ## Video description Aaron Tate, Director, Center for BrainHealth, The University of Texas at Dallas AI is the most powerful cognitive tool ever built. The question is whether you are using it to strengthen your brain or quietly letting it do the thinking for you. Research is documenting measurable neurological changes from AI use, including reduced cognitive control, eroding critical thinking, and risks to social cognition that are only beginning to be studied. But the evidence is equally clear that when AI is used intentionally, it amplifies performance. How we adopt it may matter more than whether we adopt it at all. At the Center for BrainHealth at UT Dallas, we have spent 30 years studying how the brain performs, how it declines, and how it recovers. We built validated tools to measure cognitive performance holistically, and we are now applying that same rigor to track what AI use is actually doing to brain health over time. This is not a new problem for us. Attendees will leave with a clear model for understanding how AI use patterns drive divergent cognitive outcomes and four evidence-based practices to protect and strengthen their cognitive edge. The entire tech world is focused on adopting AI faster. This talk asks the question no one is asking: what is effortless, automated thinking doing to the cognitive infrastructure everyone is running on but nobody is maintaining? [Source](https://www.youtube.com/watch?v=-Rpu7N_nK74)

Source quotes

[3:58] Dr. Sandra Chapman: Yes, it's both. [8:31] Aaron Tate: Originate before you delegate [8:31] Aaron Tate: Originate before you offload [8:31] Lee Whitmarsh: fill our visual well [8:31] Aaron Tate: Reflect wisely [13:58] Aaron Tate: you give me the first draft and then I'll do the improvement. [14:15] Aaron Tate: money talks

Source quotes

Aaron Tate: AI is the most powerful cognitive tool ever built. The question is whether you are using it to strengthen your brain or quietly letting it do the thinking for you. Aaron Tate: How we adopt it may matter more than whether we adopt it at all. [2:51] Dr. Sandra Chapman: Yes, it's both. [4:37] Aaron Tate: cognitive debt cycle [8:31] Aaron Tate: Originate Before You Delegate [8:31] Lee Whitmarsh: fill your visual well [8:31] Aaron Tate: Who did more of the effort there, me or AI? [13:58] Aaron Tate: guardrails for your own cognitive offloading [13:58] Aaron Tate: you give me the first draft and then I'll do the improvement. [13:58] Aaron Tate: money talks. Aaron Tate: what is effortless, automated thinking doing to the cognitive infrastructure everyone is running on but nobody is maintaining?

Original transcript

[0:04] So um brain health or AI can make us more  productive. It can help us be more efficient. [0:10] Um and yes, I think it can even help us be  more creative. Um but we do need to take a [0:16] second and examine like the detrimental effects  that AI can have on um our brain health. [0:23] Center for Brain Health was founded over 30  years ago by Dr. Sandra Chapman. Um we are a [0:28] translational research institution where we create  um we do research on brain health uh performance [0:36] deficits uh of all age groups and we make trans  uh we make interventional and translational um [0:46] technologies and programs to support and promote  brain health. One of those is um our kind of [0:56] flagship platform, the brain health project. And  that is the mechanism in which we deliver our [1:04] main research project. This is a 10-year  longitudinal study that is um focused on [1:11] tracking brain health and healthy adults uh  over all age groups um for 10 years at six-month [1:17] intervals. and a portion of that population  is going we're aiming to recruit a 100,000 [1:24] participants but a portion of that population  will be engaged in direct imaging um as well. [1:30] So the measure in which we use to track  brain health over that time period is the [1:36] brain health index. And the brain health index  is uh a battery of 23 separate assessments on um [1:46] behavioral and cognitive assessments that track  the factors of brain health. um during the study. [1:56] We recently published in  Nature Scientific Reports um [2:02] evidence of a uh change in brain health over a  population of 4,000 individuals that were given [2:11] uh crit uh cognitive training strategies and  shown that they had a 7 point on average. um [2:19] from 18 to 93 had an average of 7.9% increase  in their cognitive assessments as well as uh [2:27] deeper activation in the prefrontal  cortex. Um during that time [2:35] we are dedicated to not only um studying the  the brain deficits but also the brain potential [2:42] and we are moving towards a platform for  precision brain health. Um for individuals [2:51] the core of our research is really focused on  um the executive function. Then in the precon [2:57] preffrontal cortex you see um activities like  planning, reasoning, um critical thinking, your [3:06] creativity. Um all of these things are part of  what is the seat of agency in the individual. [3:15] what gives an individual the uh opportunity for  self-direction and in our study we saw that change [3:24] when we were given a training when in participants  were given a training platform that trained those [3:32] aspects of executive function that we saw that  change and that's what we call neuroplasticity. [3:37] um the ability to see the change of  activation within the brain um based on the [3:43] areas that you are exercising. That's the basic  rule of neuroplasticity. The connections that you [3:50] exercise um are strengthened and the connections  that you do not exercise are atrophied. [3:58] And so in relation to AI, I'm often reminded of  this question that Dr. Chapman gets asked. Um [4:05] I think originally it was more to do with things  like um use of cell phones and devices, but now [4:14] it's really turn towards the question regarding  AI. Is is technology good or bad for brain health? [4:21] And she always answers that question the same  way. Yes, it's both. It can be um both good for [4:27] your brain. It can be bad for your brain.  depending on how you use any technology [4:31] um will really indicate like whether or  not it has positive or negative outcomes. [4:37] See that there are definitive  like increases in productivity. Um [4:43] AI is has the ability to narrow the skills gap  between junior and senior um professionals. Um [4:50] but the continual use of AI and dependency of  AI can create a reduction in other capabilities [4:59] um your own reasoning your own critical thinking  and this is what we call the cognitive debt cycle [5:04] and it's why we have to really focus on  self- agency um as a as a remedy here and so [5:12] this cognitive debt debt cycle is defined  by when you start to use AI more regularly [5:19] um you can come more and more dependent on it,  which will um create this kind of cyclical effect [5:26] where you're depending on it more and more and  you're not exercising your own cognitive functions [5:32] and you can see a decreased capability from that.  um really interesting study of I think it was the [5:40] MIT Media Lab that was looking at a population  that had been asked to do a create a essay. Um and [5:50] the part of the population that was highly  dependent on AI um was actually creating [5:58] like structurally more sound essays um and  even had better points in creativity. But [6:05] there was a deep homogenization of um the the  concepts that were within those essays. Um and [6:13] another meas metric that came out of that was  of the ones that were heavily dependent on AI, [6:19] we saw a 83% of those were unable to give a  direct quote um or outline of their own essay. [6:28] And so you really can create a cyclical system  there that that you're not exercising the areas [6:36] of your brain that you need to to maintain your  own um cognitive function. So you can get into [6:42] the loop. So we really have two choices because  not adopting AI is really no no choice at all. um [6:49] you really can use it for cognitive ampl cognitive  amplification or you can use it for cognitive [6:57] delegation. And so if you're using AI for  cognitive amplification, what you're really [7:03] doing is you're the thought work yourself. You  are writing that rough draft. You're um doing [7:12] the initial analysis. you're um digesting the  book without asking for a summary yourself [7:19] before you go to AI. And what you can use AI for  in that context is use it for the tedious tasks, [7:25] use it for the editing, use it for um you  know the I've got the rough project plan, [7:31] I just want to add in the details, things like  that. The cognitive delegation again gets into [7:36] that cognitive debt cycle where you stop thinking  for yourself and you lose the ability to um [7:46] you can over time lose the ability to uh  have that higher order thinking. And so [7:55] we think that the differentiator in the  future will be once we all have the best AI [8:00] models right there in our operating system, it's  really going to be the people that are able to [8:08] intentionally and strategically like use AI  where it will accelerate so they can actually [8:15] do more of the thought work. if they can do more  of the higher order thinking and do more of the um [8:24] the cognitive effort then AI can  accelerate the rest of it for them. [8:31] So we're putting together some bas a basic  framework for a brain healthy adoption of AI. [8:38] So number one is basic neuroscience literacy. A  lot like um nutritional literacy and physical um [8:46] fitness literacy will help you in your  daily life make good choices towards um [8:53] you know I've probably had too many calories  today. I need to not get that dessert or [8:56] something you know along those lines. You can  make intentional decisions along the way that [9:00] will help you maintain your cognitive  control and your cognitive um exercise. [9:08] two um two is to basically originate before you  delegate um originate before you offload. So [9:18] we just want you to take the time to write that  rough draft to put in the thought work that you [9:26] need before you go to AI. And if you continually  do that, you're not offloading the core cognitive [9:32] function that um gives your brain the exercise and  workout that it needs to maintain its own health. [9:41] Three is disengage purposefully. And so on this  one, I'm I'm reminded of um a professor I had [9:50] in my undergrad. I was studying design and art  direction and she would always tell her name was [9:55] Lee Lee Whitmarsh and she would always tell us  to fill our visual well. She said as designers [10:01] you really do have to be up to date on the latest  trends, the latest movies, books. Uh she would [10:09] say visual but she would also say you know music  and just general the the the current zeitgeist of [10:16] pop culture. And she said because you need those  things so that you can respond when your client [10:24] eventually says I want it to be like this and you  know what they talk they're they're talking about. [10:29] Likewise, if you're disengaging from AI purpose  purposefully um so that you can continue to build [10:37] your own cognitive networks, you are digesting  the book before you ask for the summary. [10:42] you really can maintain your own agency over  over AI. Um and so four is reflect wisely. [10:52] This is a practice um I do is at the end of  every workday I kind of look at my task list [10:58] and what I've completed and I kind of ask myself  who did more of the effort there whether me or AI [11:04] and there's plenty of days where when I tally  everything up AI probably did more of the work [11:09] than I did. Um but you know when that becomes like  day after day I really know I need to start to [11:16] meter um how much I'm engaging with it because I  I do you know everything we do at Center for Brain [11:21] Health is about maintaining that maintaining your  cognition for the long term building resilience [11:28] now. So if there is a neurological event or  something that you have a reserve to pull from [11:35] and so anytime you're offloading too much you  need to pull back a little bit and do the work. [11:41] Um and so that is the framework that we're  putting forth but what are we doing with [11:47] what we've already done? So we have our tenure  longitudinal study. We're adding AI elements, [11:54] AI tracking measures alongside our other battery  of cognitive assessments. So we can track we'll [12:01] track things like this um the frequency of use,  the types of use, the sophistication of use of AI [12:08] and look at that in comparison to our larger brain  health index scores. We're also looking at it from [12:14] a social aspect. Right? So, a lot of the time,  um, I haven't talked a lot about the social, [12:20] um, but there's also a a social a growing social  dependency on of AI to pe to a lot of people. [12:29] Um, they're more socially isolated themselves.  So, they're leaning towards AI. We have a platform [12:35] called Charisma that has a measure of social  competencies and social cognition um, called [12:42] Isert. We're going to do the same there. We're  going to look at that those scores in relation to [12:47] the types, frequency, sophific sophistication  of use um and see how that compares. [12:55] We also uh earlier I mentioned the um cognitive  training that our participants get in the Brian [13:00] Health project. That's our smart training,  strategic memory advanced reasoning training. [13:05] um we're looking at how we can integrate uh best  practices for AI adoption, best practices for [13:11] um use there alongside our our regular  cognitive training. So because it is [13:15] something that we're just now having to you  know deal with, you know, if you look at when [13:20] we started automating all the other physical  tasks um of our life, we had to invent gyms so [13:25] that we could exert ourselves more effectively.  And same thing, we're going to add the smart um [13:31] AI best practices to our smart training.  Additionally, brain healthy AI agents. Um we're [13:38] currently working on a precision brain  health model that will help people [13:47] track and improve their own brain health. Um,  but I think there's a real opportunity there to [13:52] build in the things that we're saying are the best  practices right now into that where it's like it's [13:58] giving you um prompts back that are it's kind of  like guardrails for your own cognitive offloading. [14:04] It's prompting you to actually you give me the  first draft and then I'll do the improvement. [14:08] So build into our agents these models that  will force them to do their own cognitive work. [14:15] Um, and how do we get it more widely adopted?  Well, as with most things that money talks [14:23] and so Dr. Andrew Nevin is a economist and he has  already done a lot of research with Dr. Jackie [14:30] Gamino at Center for Brain Health on how critical  thinking skills training in middle schoolers has [14:39] shown measurable increases in their potential  earnings over the life. And he is now looking more [14:47] broadly at how adopting brain health practices  have genuine workforce, societal, economic, [14:56] and fiscal outcomes for those um organizations  that are looking to adopt those practices. And so [15:06] we're also, you know, we are a translational  research institute. Um, but we cannot, [15:14] you know, we we're part of a university and not  for-p profofit. We need organizational partners [15:20] that can take these things to market for us  with us. And so we've got these measures, [15:27] we've got the interventions. We're showing  that it not only has um potential for [15:33] individual improvements, but also for  economic and fiscal improvements. So, [15:41] thank you and please reach out. Um, there's  my email also. You can get more information at [15:48] centerforbrainhealth.org or the brainhealthpro.org  for our research study. Thank you. [applause]