Corporate AI Spending Stalls on Tool Chasing, PMI and Deans Warn

PMI与多位学者警告：盲目追逐工具致企业AI支出受阻 当机构优先考虑表面的工具推广，而非基础项目管理、结构化变革管理和持久的批判性思维时，企业级AI投资便无法带来切实的生产力提升。 专家小组成员指出，成功的AI战略在本质上是一项人才战略，并告诫企业切勿以牺牲长久的方法论和持久的批判性思维为代价，盲目追逐诸如基础提示词工程等易过时的工具操作技能。结合哥伦比亚大学、西部州长大学（WGU）、约翰斯·霍普金斯大学以及PMI等机构的实际落地经验，发言者深入探讨了浅层游戏化指标、缺乏协调的工具泛滥以及基础分析能力退化所带来的潜在风险。他们驳斥了关于初级岗位即将消亡的恐慌性论调，敦促雇主通过导师制度保护内部知识传承渠道，而非仅从外部高薪挖角。最后，他们总结了用于招聘和职业发展的实用方法，强调采用“出声思考”式候选人评估、警惕AI过度拟合的简历，并重视传统学位与可堆叠微凭证相结合的协同价值。 与基础提示词工程等表面技能相比，基础AI项目管理、稳健的数据建模、战略沟通和人才培养被严重低估。 (0:04) 若缺乏积极的变革管理框架（例如ADKAR）与刻意的习惯培养，企业AI的推广将走向失败；仅发放软件许可证只会导致低采用率与员工疏离。 (5:04) 企业必须将系统性AI素养置于盲目追逐工具的错失恐惧（FOMO）之上，因缺乏协调的工具泛滥会造成严重的组织内耗与预期错位。 (8:05) 将财务奖金与原始活跃度指标挂钩会催生表面化使用，而可持续的采纳落地则需要垂直领域的针对性培训以及团队内部核心用户的带动。 (12:13) 成功的AI工作流转型需要与一线员工协同共创，以解决实际业务痛点，并确保关键的领域专业知识始终处于落地实施的核心地位。 (15:04)

Workforce readiness in the generative AI era requires prioritizing durable methodologies, change management, and human discernment over transient software proficiencies. Panelists from enterprise and higher education argue that sustainable transformation depends on synchronizing technology rollouts with human talent development, restructuring hiring assessments, and maintaining entry-level career pipelines. A Tri-Partite Hierarchy: Perishable Versus Enduring Skills (20:14) Kathleen Walch categorizes workplace capabilities into perishable, durable, and enduring skills. Perishable skills—including tracking rapid model updates, prompt engineering syntax, and specific vendor user interfaces—are quick to learn but decay rapidly. Panelists caution that obsessively chasing individual tools exhausts organizational capacity. Instead, enterprise stability relies on durable skills with multi-year shelf lives, such as data modeling and structured project governance like the Cognitive Project Management for AI (CPMAI) framework, which guides problem scoping, data preparation, evaluation, and operational deployment. At the foundation sit enduring skills: inherently human attributes such as critical thinking, discernment, and strategic communication. Paul LaForge and Dr. Lilian Ajayi Ore emphasize that an uncritical reliance on polished AI outputs leads to cognitive atrophy or 'downskilling.' A resilient talent strategy focuses on enduring and durable capabilities, treating specific tools as adaptable instruments rather than primary qualifications. Change Management and Incentives Over Tool Provisioning (5:04) Deploying software licenses without structured enablement consistently results in low adoption, as demonstrated during enterprise rollouts of tools like Microsoft 365 Copilot. Drawing on Prosci's ADKAR change management framework, Walch notes that employees bypass new systems unless organizations actively cultivate an 'AI-first' habit. Tool proliferation without fundamental education can also cause organizational fragmentation, exemplified by Western Governors University discovering at least 76 uncoordinated AI initiatives across departments. Overcoming friction requires cultural safety and aligned incentives rather than punitive mandates or shallow usage tracking. At Columbia University, an anonymous baseline readiness survey removed the stigma around technical skill gaps. Furthermore, panelists warn that offering financial bonuses tied to software activity logs merely encourages superficial compliance, such as using Copilot solely to rewrite a morning email. Effective adoption instead leverages embedded peer power users and co-creation projects—such as WGU's 'anti-Control-F' initiative—where frontline staff automate repetitive tasks using their own domain expertise. Candidate Fluency and the Junior Talent Paradox (28:53) As job candidates increasingly use generative AI to produce 'overfit' resumes that match job descriptions mathematically, conventional screening mechanisms fail. To evaluate authentic fluency, LaForge recommends think-aloud interview exercises where applicants explain trade-offs and diagnose flaws in generated code in real time. Dr. Ore and Walch advocate using structured, scenario-based inquiries based on CPMAI phases, requiring candidates to articulate where a project started, what errors occurred, and how they applied ethical discernment to decide whether a technically feasible solution should actually be deployed. This rigorous evaluation is particularly vital when assessing early-career talent. While digital natives navigate software interfaces effortlessly, LaForge points out that empirical observations reveal a lack of true AI literacy: younger workers frequently place uncritical trust in automated outputs. True seniority requires the enduring skepticism to question and verify AI-generated answers, making verification skills the true differentiator between junior output and senior capability. Preserving Talent Pipelines Through Stackable Credentials (23:08) Macroeconomic pressures and AI integration have prompted many employers to freeze entry-level hiring and prioritize experienced personnel. Dr. Ore warns that exclusively hiring senior talent risks depleting the institutional knowledge necessary for long-term organizational viability. To bridge this gap, universities and employers are shifting from passive lecture formats toward hands-on apprenticeships, guided mentorships, and programs designed to accelerate junior practitioners toward senior-level competency. Simultaneously, workforce education is moving toward modular credentialing. Aisha Lawrey and Kathleen Walch describe partnerships between academic institutions and professional bodies, such as integrating Project Management Institute (PMI) standards directly into degree programs so students graduate with both academic diplomas and credentials like the PMP. Expanding pathways such as the Certified Associate in Project Management (CAPM) into high schools provides early occupational access, reinforcing that future career development is an additive blend of formal degrees and targeted micro-certifications. While organizational training budgets are expanding, an unresolved challenge remains in synchronizing rapid technical deployments with human readiness to avoid black-box dependence and skill degradation. Panelists emphasize that unless organizations preserve entry-level pathways and cultivate intrinsic, cross-disciplinary problem solving, investments in advanced technical infrastructure will fail to produce sustained business value.

在Ai4研讨会上，来自项目管理协会（PMI）、西部州长大学（WGU）、哥伦比亚大学及约翰斯·霍普金斯大学的专家指出，企业在AI转型中过度押注于易过期的工具表层操作，而严重忽视了底层数据建模与人才素养建设。专家主张，构建未来劳动力蓝图必须依托结构化变革管理、业务一线协同共创及产学深度融合，以培育抗周期的持久与长青技能，化解技术断层和初级岗位萎缩的风险。 超越工具追逐：重塑三层技能生命周期 (20:14) 小组专家一致认为，当前劳动力市场对AI技能的估值存在严重偏差。PMI的Kathleen Walch与WGU的Paul LaForge指出，基础提示词工程与直接生成工具产出属于最被高估的能力，而真正的短板在于从业务问题定义出发的AI项目全流程管理，以及支撑大语言模型运行的底层数据建模。约翰斯·霍普金斯大学的Aisha Lawrey与哥伦比亚大学的Lilian Ajayi Ore进一步补充，单纯投向硬件基础设施的技术战略若脱离了组织沟通与人才发展，将无法释放应有价值。 针对AI技能的迭代风险，Walch提出了“易腐技能”、“持久技能”与“长青技能”的三层框架。特定工具操作与提示词语法属于快速贬值的易腐技能，过度追逐易引发组织疲劳；以CPMAI（人工智能认知项目管理）六阶段方法论（从业务理解、数据准备到部署评估）为代表的系统设计属于稳步演进的持久技能；而批判性思维、甄别力与战略沟通则是永久不可替代的长青能力。专家强调，专业人员应以持久与长青技能为支柱，避免在频繁更迭的工具表象中迷失方向。 结构化变革管理：破除形式主义培训与浅层指标 (5:04) 工具的大规模铺开如果缺乏系统性素养培育，往往会引发组织混乱。LaForge透露，WGU曾在各部门盲目并发运行了76个以上的去中心化AI项目，最终因工具碎片化和基础认知脱节而被迫调整战略。Ajayi Ore分享哥伦比亚大学的实践指出，面对教职人员对岗位替代的焦虑及对技术短板的防御心理，校方通过全匿名的准备度基线调研消除了能力差距带来的羞耻感，从而平稳推进了全校范围内的AI素养课程。 在落地推广中，Walch强调必须借助Prosci与ADKAR变革管理模型，将工具使用固化为行为习惯。若仅以工具使用频率等浅层数据作为绩效考核，例如将3%的加薪与日常AI操作绑定，员工只会为了完成指标而让Copilot润色早间邮件，造成严重的资源浪费。专家建议，企业应针对财务、营销等具体业务垂直场景设计差异化培训，并在团队内设立常驻超级用户，以在软件界面高频调整时提供及时的点对点支持。 业务共创与认知摩擦：遏制专业技能退化 (15:04) 解决技术落地受阻的关键在于吸纳一线业务人员参与流程共创。LaForge介绍，WGU员工因厌倦日常繁琐的手工检索而自发启动了“反Control-F”行动，业务骨干在协同设计AI自动化流程的过程中自然习得了Python编程能力。专家强调，任何脱离实际业务专家的自上而下的AI改造都极易落空，因为核心领域认知始终掌握在岗位人员手中。 针对安永近期报告所反映的“AI投入激增但落地成效低迷”的生产力悖论，Ajayi Ore解释称，这源于技术铺设速度大幅超越了组织承载力。过度依赖黑箱系统还会引发稀缺专业能力退化（即“降能”现象）。专家呼吁在工作流中必须保留刻意的“认知摩擦”，专业人员需像“不借助计算器进行算术验算”一样，定期手动执行任务以保持核心批判能力。此外，讨论还将视角延展至汽车技师、电工等蓝领技能工种，指出具备诊断辅助能力的AI同样正在重塑实体工种的认证与作业形态。 重构就业通道：防范简历拟合与初级招聘冻结 (23:08) 在宏观经济不确定性与技术颠覆的双重压力下，企业倾向于冻结初级岗位、转向高薪挖掘资深人才。Ajayi Ore依据其WIN模型呼吁雇主切勿一味“自顶向下”挖角，否则将掏空企业长期生存所倚重的组织知识积淀。为了应对雇主对即战力的苛刻要求，高等教育必须从被动讲授转向基于导师制的实操学徒模式，承担起“让初级毕业生具备高级产出能力”的培养任务。 面对求职者利用生成式AI使简历过度拟合职位的现象，LaForge与Walch建议面试官摒弃程式化问答，转而采用“出声思考”解题及CPMAI六阶段穿透式追问，重点考察候选人在代码异味识别、错误排查及技术伦理权衡方面的真实推导过程。与此同时，高校与行业协会正积极推动堆叠式微凭证体系，如PMI将CAPM初级认证前置推广至高中阶段（年满18岁即可报考），并将PMP资质嵌入学位课程，倡导将传统学位与敏捷职业认证深度咬合的“兼收并蓄”发展路径。 该研讨会明确指出，AI应当被定位于人类专业能力的放大器而非替代物。尽管微凭证与产学实操合作能够部分弥合就业门槛升高的鸿沟，但在企业普遍追求即时回报的功利考核导向下，如何建立跨行业的标准化质量评估、长期防范生成式工具对员工批判性求证能力的侵蚀，仍是悬而未决的系统性挑战。

## Background 本视频由 Ai4 于 2026 年 8 月 22 日发布在 YouTube 平台上。内容记录了由《财富》（Fortune）杂志编辑总监 Kristin Stoller 主持的圆桌论坛，汇聚了来自项目管理协会（PMI）、西部州长大学（WGU）、哥伦比亚大学及约翰斯·霍普金斯大学的行业与学术领袖。论坛聚焦于企业如何跳出单纯采购软件的技术思维，通过自上而下的员工技能提升、重塑岗位职责以及构建“以人为本”的人才战略，将劳动力就绪度转化为 AI 时代的竞争壁垒。 ## TL;DR 论坛深入探讨了 AI 浪潮下人才战略与组织转型的核心逻辑，指出基础 Prompt 工程与单纯堆砌技术被严重高估，而项目管理、底层数据素养和战略沟通才是真正的核心竞争力。专家们强调，缺乏培训与变革管理的工具下发以及浅层的“打勾式”奖金考核往往导致形式主义，企业必须通过业务共创、部门定制化赋能以及超级用户网络来扎实推进 AI 素养。基于“易损技能、耐用技能与持久技能”的三层技能模型，嘉宾们呼吁高校通过现代学徒制和微认证破解初级岗位收缩困境，并分享了如何在招聘面试中利用“出声思考”和多案例追问来穿透 AI 过度拟合的简历、考察求职者真实的批判性思维。 ## Atmosphere & style 整场圆桌对谈氛围务实、敏锐且充满思辨活力，在主持人的利落串联下展现出产业界与学术界深度碰撞的默契。嘉宾们毫不避讳地坦陈组织在早期创新过载、盲目追逐工具中的真实教训，并用“甜品餐盘”等生动比喻化解复杂的技能分类。讨论在直面青年就业焦虑与技能退化风险的同时，始终保持着建设性与温度，传递出拒绝浮躁追风、回归底层人类价值与终身学习的从容信念。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=4s) 在 AI 转型中，单纯的技术战略与基础 Prompt 工程常被高估，而底层数据质量、业务问题导向的 AI 项目管理以及人才战略才是最被低估的核心能力。 - [05:06](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=306s) 企业在推广 AI 工具时切忌“只给工具不给培训”，必须借助 ADKAR 变革管理模型建立员工的认知（Awareness）与意愿（Desire），优先培养 AI 素养而非盲目堆砌工具。 - [10:58](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=658s) 将 AI 使用量与绩效奖金挂钩容易引发“每天仅用 AI 重写一封邮件应付打勾”的浅层形式主义；有效的落地方式是按部门垂直定制培训并建立核心用户（Power Users）互助网络。 - [17:03](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1023s) 高校与企业合作表明，雇主并不看重求职者具体掌握哪款特定工具，而是渴求具备批判性思维与解决实际业务问题能力的人才。 - [20:15](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1215s) 技能体系可划分为易损技能（特定工具操作）、耐用技能（如 CPMAI 项目管理框架）与长青/持久技能（批判性思维与洞察力），职场人需保持均衡，避免陷入盲目追逐新工具的疲惫陷阱。 - [23:08](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1388s) 入门级岗位收缩主要源于宏观不确定性，高等教育亟需转型为融入 AI 协作的“现代学徒制”，教会初级人才以资深工程师的素养与成熟度走出校门。 - [25:25](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1525s) 企业若完全关闭初级岗位招聘通道，将导致人才梯队断层并严重流失维系长期发展的组织机构知识（Institutional Knowledge）。 - [28:52](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1732s) 为应对简历被 AI“过度拟合”与虚饰问题，面试应采用“出声思考”（Think Aloud）与双业务案例深度追问，重点考察候选人的独立贡献、纠错过程与伦理判断力。 - [33:24](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2004s) 年轻一代虽为“工具达人”，但普遍缺乏 AI 素养与质疑精神；只有同时兼备工具熟练度、批判性审视与底线判断力，才能成长为真正的资深人才。 - [34:50](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2090s) 防范因依赖 AI 导致技能退化（Downskilling）的关键在于“流程同步”：保持人机协作节奏一致、坚持批判性思维日常训练，并在适当时机回归传统手工操作以保持思维敏锐度。 - [42:31](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2551s) 传统学位与微证书（如 PMP、CAPM）并非对立关系，而是互补的“两者皆可”终身学习工具箱，职场人可根据专项技能缺口进行精准自我投资。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=4s) 开场破冰：未来五年最被高估与低估的技能 论坛主持人是来自《财富》（Fortune）杂志的编辑总监 Kristin Stoller。她负责报道工作未来（Future of Work）议题并撰写周一专栏。在开场环节，Kristin 向四位嘉宾提出了一个快速破冰问题：用一分钟左右介绍自己的姓名、职位、机构，并指出“未来五年内哪项技能最被高估，哪项技能最被低估”。 ```mermaid flowchart TD accTitle: Panelist views on AI skills subgraph Panelists["嘉宾与核心观点"] K["Kathleen Walch (PMI)"] P["Paul LaForge (WGU)"] L["Lilian Ajayi Ore (哥伦比亚大学)"] A["Aisha Lawrey (约翰斯·霍普金斯大学)"] end subgraph Overrated["最被高估的技能/策略"] O1["基础 Prompt 工程"] O2["仅用 AI 工具生成结果"] O3["单纯的技术战略 / 基础设施堆砌"] end subgraph Underrated["最被低估的技能/策略"] U1["AI 项目管理与以业务问题为起点 (CPMAI)"] U2["底层数据模型与数据质量"] U3["战略沟通与学习心态 (Openness)"] U4["人才战略 (Talent Strategy)"] end K --> O1 K --> U1 P --> O2 P --> U2 L --> U3 A --> O3 A --> U4 ``` Project Management Institute（项目管理协会，PMI）的人工智能参与与社区总监 Kathleen Walch 首先发言。当天恰逢她加入 PMI 两周年——她此前联合创办的公司被 PMI 收购，并带来了管理 AI 项目的 CPMAI 认证（目前是 PMI 增长最快的认证），同时她主持的播客《AI Today》也已播至第 10 季。Kathleen 认为最被低估的技能是“根本性理解如何运行和管理 AI 项目”，尤其是认识到底层数据的重要性，以及从要解决的业务问题出发，而不是直接跳到第四阶段（模型开发或 Agent 开发）；最被高估的技能则是“基础 Prompt 工程（Basic Prompt Engineering）”。 Western Governors University（西部州长大学，WGU）技术学院副总裁兼院长 Paul LaForge 负责管理计算机科学、软件工程、AI 和数据等专业。他认为最被高估的技能是“仅仅能够使用 AI 工具生成输出结果”，而最被低估的则是“数据相关技能”。他强调人们往往忽略了数据，但只有在 LLM 底层拥有坚实的数据模型时，AI 的表现才会出色。 哥伦比亚大学专业研究学院（SPS）全职讲师、AI 核心导师 Dr. Lilian Ajayi Ore（同时也是非营利组织 Global Connection for Women 的创办人）指出，最被低估的技能是“战略沟通（Strategic Communication）”——即如何表达采纳 AI 的价值和必要性，并向全体员工清晰沟通，防止有人在 AI 加速发展中掉队；最具价值的特质则是“开放的学习心态（Openness / Learning Mindset）”，即愿意主动学习并接受职场中的变革。 约翰斯·霍普金斯大学 Whiting 工程学院的高管与专业教育部门负责任 AI 组合负责人 Aisha Lawrey 认为，最被低估的是“人才战略（Talent Strategy）”及培养人才的相关能力，最被高估的则是单纯的“技术战略（Technology Strategy）”。企业往往在搭建基础设施上投入过多精力，却在负责管理和运营这些设施的人才培养上投入不足。 --- ### [05:06](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=306s) 现场互动与企业重塑技能案例：工具派发 vs 培训先行 Kristin 向现场观众提问以活跃气氛，询问有多少人所在的企业今年在“学习与发展（L&D）”上的投入超过去年。现场大部分观众举手，且无人举手表示减少投入。随后，Kristin 引导嘉宾分享企业在员工技能提升（Upskilling）过程中的真实案例：哪些尝试失败了，哪些方案最终取得了成功。 Kathleen Walch 坦言，从自主创业当老板转换到大型组织是一次巨大的文化冲击。谈及 AI 时，变革管理（Change Management）必不可少，为此她专门考取了 Prosci 变革管理认证，以便用专业框架理解这一转型。她指出，**“绝对不能在没有培训的情况下直接把工具丢给员工”**。如果仅仅向员工开放权限，由于缺乏 ADKAR 模型中的“认知（Awareness）”和“意愿（Desire，即了解对自己有何好处）”，员工在日常繁忙中根本不会使用。PMI 是一家 Microsoft 体系组织，全员部署了 Microsoft Copilot。PMI 并没有随意下发工具，而是投资制定了完整的推广方案与系统培训，推动员工形成“AI 优先、绝不从空白页面起步（AI-first, never start with a blank page）”的工作肌肉记忆，从而取得了显著成效。 - **现场调研结果**：现场多数参会企业今年在 L&D 方面的预算和投入均高于去年，无人缩减。 - **ADKAR 变革模型落地**：工具推广若缺乏 Awareness（认知）和 Desire（意愿），采纳率必然低下。 --- ### [08:06](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=486s) WGU 的创新过载与哥伦比亚大学的 AI 素养破冰 Paul LaForge 分享了 WGU 的教训：WGU 高度注重创新，全校曾一度并行推进了至少 76 个不同的 AI 项目（甚至可能更多，只是后来停止了计数）。然而，在这种蜂拥而上的创新热潮与工具曝光下，“AI 素养（AI Literacy）”却被远远甩在后面。WGU 不得不掉头补课。Paul 反思道，如果可以重来并减少“错失恐惧（FOMO）”，他们一定会**先抓 AI 素养建设，然后再跟进工具的引入**。 哥伦比亚大学的 Dr. Lilian Ajayi Ore 则分享了高校场景下的成功落地实践。哥伦比亚大学在允许学生公开使用 AI 方面起步相对审慎。在获批全校推广时，团队敏锐地注意到了师生之间的信任危机——调研报告显示，许多学生感到教师对他们抱有敌意，甚至在学生未用 AI 时也无端怀疑其作弊；而教师群体内部也充斥着“AI 是否会取代我的教学工作”的焦虑。 Lilian 团队与 IT 部门合作，专为教职员工设计了一套定制化的 AI 素养培训体系： 1. **重塑认知定位**：向教师明确阐明，培训目的不是惩罚或暴露谁不会用 AI，而是提供教学提效的工具。 2. **匿名调研诊断学习缺口**：通过匿名问卷收集教师的准备度与真实痛点，彻底消除因不懂 AI 而产生的羞耻感与病态标签化（Stigmatization）。 3. **分层推进采纳**：从基础素养培训切入，打通心理屏障后，教师不仅大幅降低了抵触情绪，反而对 AI 的教学辅助价值产生了由衷的认可与主动应用。 ### [10:58](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=658s) 推广 AI 素养与激励机制的陷阱：超越“打勾式”奖金 在讨论如何推动全员接受 AI 时，讲者们强调了循序渐进与自愿原则的重要性。培训不应被强加于人，而应作为助力个人成长的赋能工具。哥伦比亚大学等机构在推行培训时，并未将其设为强制性课程，而是作为一种“强烈推荐”的专业发展工具，这让教职人员更容易接受，同时也帮助他们掌握了与学生探讨 AI 的共同语言。约翰斯·霍普金斯大学的 Aisha Lawrey 补充指出，必须清晰区分“AI 工具（AI Tools）”与“AI 素养（AI Literacy）”。许多机构过度把精力耗费在具体技术上，而忽略了基础教育，这往往会导致员工产生抵触心理，或者对 AI 的真实能力产生不切实际的高估。 针对如何激励员工持续进行终身学习，主持人 Kristin Stoller 提出了企业界的普遍做法：部分公司甚至将 AI 使用情况与绩效奖金挂钩（例如使用 AI 即可获得 3% 的加薪或奖金）。然而，项目管理协会（PMI）的 Kathleen Walch 警告称，这种奖金激励存在巨大的陷阱。一旦将衡量标准单纯设为“AI 使用量”，很容易催生出“高频却浅层（heavy usage that's very shallow）”的形式主义。以企业部署的 Microsoft 365 Copilot 为例，系统可以追踪员工在 Chat、邮件、PowerPoint 或 Excel 中的使用频次；如果每天仅用 AI 重写一封邮件就能被判定为“活跃用户”，员工为了拿到奖金，每天上班的第一件事就会是“用 AI 重新生成一封邮件，打上勾，拿走奖金并继续干别的”，这完全背离了引入 AI 的初衷。 ```mermaid flowchart TD accTitle: 浅层激励与深度赋能模式对比 subgraph Bad["浅层指标导向（容易失效）"] B1["设定 KPI / 3% 奖金挂钩"] --> B2["员工浅层应付<br/>（每日仅用 AI 重写一封邮件）"] B2 --> B3["系统完成'打勾'<br/>（未解决实际业务痛点）"] end subgraph Good["垂直赋能与专家网络（有效落地）"] G1["挖掘员工切身利益<br/>（'What's in it for them'）"] --> G2["按部门垂直定制培训<br/>（财务、市场、教学各取所需）"] G2 --> G3["建立 Copilot 核心用户网络<br/>（5 分钟即时解惑，化解界面变动挫败感）"] end ``` 为了真正激发内生动力，管理者必须向员工讲清“这对我有什么好处（what's in it for them）”。Kathleen 建议采取更精准的垂直化策略： - **按部门定制培训**：在完成通用基础培训后，应将组织按职能细分（例如财务部、市场部、教学研发团队），分别针对其特定业务场景提供定制化培训。 - **设立 Copilot 核心用户（Power Users / Champions）**：AI 工具更迭极快（尤其是微软等厂商频繁改动按钮和界面布局），几个月前的教学视频很快就会失效。通过在各团队培养深度的“超级用户”，员工遇到卡点时只需进行 5 分钟的简短交流，就能省去数小时的摸索与挫败感。 --- ### [15:05](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=905s) 业务共创与全天候支持：WGU 的“反 Ctrl+F”实践 西部州立大学（WGU）副校长兼技术学院院长 Paul LaForge 分享了如何通过“共创模式（Co-creation Model）”消除员工对 AI 取代工作的恐惧。他指出，许多员工对岗位中的某些枯燥环节深恶痛绝，极其渴望摆脱这些繁琐任务。WGU 曾与一个团队合作，发现他们日常工作中最讨厌的就是频繁按快捷键查找信息，于是学校团队便顺势发起了所谓的“反 Ctrl+F 运动”（anti-control F）。 Paul 强调，企业在推进 AI 转型时最容易犯的错误，就是脱离一线业务人员的专业背景。那些具备深厚领域专业知识（Domain Expertise）的员工正坐在工位上，如果脱离他们去设计解决方案，AI 转型必然会走向失败。通过让这些员工参与到方案的共创中，员工会主动发现：“我以前懂一点 Python，但过去总在机械地按 Ctrl+F；现在参与到 AI 方案构建中，我自然而然地完成了技能重塑（Upskilling）。” 而在学生端，WGU 则采取了灵活的自愿使用机制（Opt-in）： - **全天候个性化辅导**：学生可以随时利用面向学生的 AI 工具进行模拟测验、随堂练习，或者就学业财务等焦虑问题寻求建议。 - **填补深夜服务空白**：WGU 许多学生习惯在凌晨 3 点（WGU 典型的学习时间段）上线学习。在无法拨打人工客服电话的深夜，AI 助手提供了宝贵的即时支持与陪伴。 --- ### [17:03](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1023s) 衡量提升成效：企业雇主看重的人才特质与跨学科协作 在探讨如何衡量员工或学生 AI 技能提升的成效时，哥伦比亚大学的 Lilian Ajayi Ore 指出，传统的课程评估表（Course Evaluation）与定期的教学研讨会（Pedagogical Lab）是直接的衡量指标，但更核心的是评估受训者是否做好了进入职场的准备。在与校董会成员及企业雇主面对面沟通“你们到底在招募什么样的人才”时，企业给出了明确的答复：他们对学生在校期间具体学会了哪一款特定软件并不太看重，他们真正渴求的是**问题解决者（Problem Solvers）**与**具备批判性思维的人（Critical Thinkers）**——即那些能够清晰阐述如何综合利用传统工具与 AI 工具来解决实际业务难题的人。 ```mermaid flowchart LR accTitle: 技能重塑与跨学科协作闭环 A["企业核心诉求<br/>（批判性思维 & 解决问题能力）"] --> B["教学与培训机制<br/>（开放论坛、教学实验室）"] B --> C["打破部门孤岛<br/>（跨学科协作：工程 + 教学 + 设计）"] C --> D["建立健康学习文化<br/>（徽章认可而非过度金钱激励）"] D --> A ``` 关于激励机制的平衡，Lilian 提出了深刻的见解：过度的外部物质激励会扭曲学习初衷，把文化变成被动的“为了拿奖励而应付差事”。组织应当构建的是一种将“学习成果本身作为回报”的学习文化，辅以对早期采纳者（Early Adopters）和推广大使授予“特殊徽章”等荣誉认可，而非堆砌过多物质刺激。 Aisha Lawrey 则从跨界协作的角度补充了衡量标准。她认为评估不应仅看课程完成率，还要看组织是否为员工提供了跨部门探索的自由度。过去各院系和部门往往各自为政（如工程学院与教育学院、设计学院之间存在壁垒）；当给教师和学生打破孤岛（Silos）、展开跨职能合作（Cross-functional Collaboration）的自主权时，探索新领域本身就变成了极具吸引力的内在激励，同时也大幅缓解了岗位可能被淘汰的技术焦虑。 --- ### [20:15](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1215s) 面向未来的技能框架：易损技能（Perishable Skills）与避免盲目追逐工具 针对行业中频繁出现的“批判性思维”、“商业判断力”等热门词汇，Kathleen Walch 提出了一个系统性的技能分类框架，将人类在 AI 时代需要掌握的技能划分为三大类别：**易损技能（Perishable Skills）**、**耐用技能（Durable Skills）**与**长青技能（Enduring Skills）**。 在谈及第一类“易损技能”时，Kathleen 阐释了其核心特征与应对原则： - **易损技能的定义**：学习速度快，但失效和退化的速度同样极快（Quick to learn and quick to decay）。 - **具体表现**：紧跟各大主流大模型（LLM）、前沿模型（Frontier Models）及各类 AI 工具提供商几乎每天发布的最新动态、功能更新或界面变化。 - **避免盲目追逐工具（No Tool Chasing）**：尽管了解并适应工具变化是必要的，但个人和组织切忌盲目跟风追逐市面上的每一款新工具。企业应当立足于自身现有的技术生态——如果是微软生态体系（Microsoft Shop），就深耕 Copilot；如果是谷歌生态体系（Google Shop），就专注于对应的谷歌 AI 工具，将有限的精力投入到核心业务场景中。 ### [21:15](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1275s) 技能的“餐盘理论”：易损、耐用与持久技能的平衡 Kathleen Walch 深入剖析了人工智能时代下人才技能的三层架构，并警告职场人切忌陷入无休止追逐单一工具的疲惫陷阱。如果你试图把市面上层出不穷的工具（例如刚学会 Gemini 又立刻转向其他新模型）全部学个遍，“你八成会彻底疯掉，甚至可能已经开始吃不消了”。她将这些具体的工具操作定义为“易损技能”（Perishable Skills），它们更迭迅速且生命周期极短。与之相对的是“耐用技能”（Durable Skills），这类技能学习周期较长，但保质期也显著更久，主要涵盖系统性框架与方法论，例如用于规划和管理全流程 AI 项目的 CPMAI（Cognitive Project Management for AI）方法论——它指导团队遵循严密的递进步骤：明确业务痛点、评估数据需求、构建模型或智能体（Agent）、展开严格测试，直至最终投入现实生产。 在架构的最底层，是价值最为恒定的“持久技能”（Enduring Skills），即批判性思维、跨部门协作、沟通表达、洞察力（Discernment）与审慎判断力。这些软技能的沉淀耗时最长，却具备极高的跨岗位迁移性。Kathleen 用生动的“自助餐盘”比喻来阐释三者的有机平衡：如果任由孩子奔向自助餐厅，他们肯定会直接冲向甜品台，大嚼巧克力蛋糕、布朗尼和棒棒糖蛋糕，随后就会遭遇血糖崩盘（Sugar Crash），不得不永远处于被动追逐下一餐的饥渴中——这恰恰对应了只盲目追捧易损技能的偏颇做法。 ```mermaid flowchart TD accTitle: 三层技能餐盘模型 accDescription: 展示易损技能、耐用技能与持久技能之间的层次与依赖关系。 subgraph Plate ["全面均衡的 AI 技能餐盘"] C["持久技能 (Enduring Skills)<br/>• 批判性思维与审慎判断<br/>• 深度沟通、协作与业务洞察"] B["耐用技能 (Durable Skills)<br/>• CPMAI 等工程与管理方法论<br/>• 业务需求向技术方案的转化"] A["易损技能 (Perishable Skills)<br/>• 特定工具与平台操作 (如 Gemini)<br/>• 易过期的短期技术技巧"] C --> B B --> A end ``` 企业在招聘中之所以更加看重持久技能，正是因为掌握了扎实批判性思维与判断力的员工，面对不断轮替的技术工具能够快速触类旁通、自主上手。 - **易损技能（Perishable Skills）**：针对具体工具或界面的操作技能，半衰期短，技术迭代会导致其快速贬值。 - **耐用技能（Durable Skills）**：项目管理逻辑与工程实施框架（如 CPMAI 规范：定义业务痛点 $\rightarrow$ 数据准备 $\rightarrow$ 模型/智能体开发 $\rightarrow$ 验证测试 $\rightarrow$ 上线运维）。 - **持久技能（Enduring Skills）**：难以被技术替代的通用内核，包含在复杂局势中去伪存真的辨析力与决策力。 --- ### [23:08](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1388s) 破解入门级岗位消亡论：高等教育与学徒制转型 主持人 Kristin Stoller 将话题引向全场最为关切的青年就业痛点。外界舆论充斥着耸人听闻的标题，断言“Z 世代与 Alpha 世代正坠入深渊”、“入门级职业生涯已彻底死亡”，但转头又出现相反论调，声称岗位并未消失、只是形式发生了转变。面对这种每周都在翻转的混乱叙事，来自学术一线的高校领袖们给出了严谨的实情剖析。 Western Governors University（WGU）技术学院副总裁兼院长 Paul LaForge 直言，真实世界绝非非黑即白的单向因果，而是多重力量交织的结果。当前初级岗位收缩的首要诱因是宏观商业环境的高度不确定性。无论是面对前沿 AI 技术的颠覆冲击，还是复杂的宏观经济波动，当企业充满顾虑时，其本能的避险策略往往是缩减初级员工招聘，转向争抢拿来即用的资深人才（Seniors）。为此，WGU 正在通过深度的雇主伙伴关系，推出针对性的 AI 工程项目，协助企业完成在职转型与技能升级，缓解用工恐慌。 更深刻的变革发生在高校内部。Paul 指出：“坦白说，高等教育自身必须被重构。”院校必须从传统的纯理论授课，全面转向融入体验式学习的“现代学徒制”（Apprenticeship Model）。在传统路径下，初级开发者往往需要漫长的磨炼，而如今帮助初学者迅速达到高阶水准的最有效路径，正是由资深导师带着学生与 AI 协同工作。教育机构的核心使命必须演变为：“教会一名初级新手，如何以资深工程师的素养与成熟度走出校门。” --- ### [25:25](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1525s) 守护人才输送线：行业协同、微认证与组织知识传承 哥伦比亚大学专业研究学院（SPS）AI 核心教职成员 Lilian Ajayi Ore 从高校办学的责任感切入探讨。攻读硕士学位的学生在经济、心理和情感上付出了巨大投入，院校必须具备极强的紧迫感，确保知识传递精准贴合产业实际，使毕业生能在毕业后 4 到 6 个月的窗口期内迅速确立竞争优势，而非耗费一年时间苦等机会。哥大引入了严格的“质量至上”（Quality Matters）评估模型，从教学法、课堂模拟、求职训练到产业匹配度展开全方位品控。 Lilian 进而向企业雇主发出强烈呼吁：切勿在恐慌中关闭初级与中级岗位的通道。求职者的技术背景极具多样性，有人带着两三年从业经验返校进修，也有人因本科就业市场严峻而直升硕士。如果所有企业都只图现成、“一律只从顶层招揽资深人才”，必将导致产业人才梯队出现断层，进而抽空维系企业未来数十年基业所必需的“组织机构知识”（Institutional Knowledge）。 约翰斯·霍普金斯大学（Johns Hopkins University）工程学院负责负责任 AI 项目的 Aisha Lawrey 进一步补充了破局手段。除了大力推行行业导师进课堂的“实践教授”（Professors of Practice）制度外，微认证（Micro-credentialing）与权威行业证书已成为企业选拔人才时的硬通货。高校正在主动协助学生遴选与考取最具产业认可度的专业认证，帮助学子在激烈的竞争中获得更强的议价能力与职业机动性。 - **学业回报时效**：通过紧密对接产业需求，力争将研究生的求职破局周期压缩在毕业后 4 至 6 个月内。 - **教学质量管控**：推行 Quality Matters 质量评估标准，横向覆盖教学法、课堂实操及就业市场匹配度。 - **组织知识保护机制**：若用人单位彻底斩断初级与中级职位的招聘入口，将会直接诱发代际传承断层与组织机构知识的枯竭。 - **复合凭证价值**：强调学位学历与微认证、专业资质证书的深度绑定，强化求职者的灵活性与辨识度。 --- ### [28:52](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1732s) 现场互动：如何在招聘面试中精准考察 AI 素养与批判性思维 在随后的现场问答环节，一位现场观众提出了一个令所有招聘方困扰的难题：在面试环节中，究竟该提出怎样的问题，才能真正撬开候选人的话匣子，有效穿透表象来评估其真实的 AI 流畅度（AI Fluency）与 AI 素养（AI Literacy）？ ```mermaid flowchart TD accTitle: 候选人 AI 素养深度评估流程 accDescription: 展示技术考查与业务案例双轨并行的 AI 面试评估框架。 subgraph Eval ["AI 素养全维度考察路径"] Q["面试考察入口"] --> T["技术层：出声思考 (Think Aloud)"] Q --> B["业务层：双真实案例追问 (Consistency Check)"] T --> T1["代码与架构评审"] T --> T2["捕捉坏味道 (Code Smells)"] T --> T3["剖析提示词逻辑与'为什么'"] B --> B1["业务痛点与技术选型"] B --> B2["界定个人独立贡献 (非 AI 代劳)"] B --> B3["复述第二案例验证一致性"] end ``` 面对这一问题，嘉宾们从技术实操与业务落地两个维度提供了破局策略： Paul LaForge 主张采用极客式的“出声思考”（Think-aloud Protocol）方法。无论让候选人编写提示词还是现场检视代码，关键在于要求他们实时讲出脑海中的推演过程：“我正在考虑什么、我认为眼前这段输出存在什么隐患”。尤其在技术研发场景下，可以给出一段包含潜在缺陷的架构或代码，让候选人主动找出其中的“代码坏味道”（Bad Smells）并说明优化理由。考官借此捕捉到的不仅是答题结果，更是其批判性思维背后的核心依据（The why behind the critical thinking）与审美判断力。 Lilian Ajayi Ore 则从业务实战角度切入。鉴于当前求职者普遍使用 AI 润色甚至代写简历，用人单位的核心诉求是甄别候选人脱离了 AI 辅助到底是否具备真实交付能力。她建议面试官直接抛出具体的业务案例追问： 1. 你们过去的业务遇到了什么具体挑战？ 2. 你是如何设计并落地 AI 战略来解决该问题的？ 3. 过程中使用了哪些自研或商用工具？ 4. **你个人的独立贡献到底是什么？** Lilian 特别强调，必须让候选人列举两个独立的具体案例，以此交叉核验其阐述故事的前后连贯性与真实度，进而考察其往往被低估的关键能力——清晰向用人单位传达自身业务价值的高效沟通力。 紧接着，Kathleen Walch 展开了她的核心视角：企业必须分清两层截然不同的评估维度——其一是“如何利用 AI 工具实现个体降本增效”，其二则是“如何以全局视角规范化地运营与管理 AI 项目”。项目管理协会（PMI）在实践中格外注重后者，而这一能力绝非单纯会写两句 Prompt 就能涵盖。 ### [31:32](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1892s) AI 过度拟合简历的隐患与真实思维展示 在招聘场景中，候选人过度利用 AI 生成简历已成为令人头疼的问题。Kathleen Walch 借用机器学习中的概念指出，现在很多简历出现了严重的“过度拟合”（Overfitting）现象——内容被修改得与招聘要求严丝合缝，但一核实就会发现求职者根本不具备这些技能。Kathleen 分享了自己的真实经历：HR 部门曾推荐了两位看似条件完美的候选人进入面试，但她一看简历就发现内容定制得过于刻意，实际上完全不合格，她不得不专门向 HR 团队提出预警。这种做法会让求职者失去“真实的自我”（authentic human self）；如果企业只想要机械匹配的产物，完全可以直接创建 AI 智能体（AI agent）。 为了考察求职者真实的思考过程与解决问题的能力，Kathleen 在面试中会要求候选人详细拆解项目管理流程（例如 CPMAI 方法论）： - 阐述其在管理和运行 AI 项目时具体如何应用该方法论。 - 考察求职者从哪一个阶段起步（整个框架共六个阶段，正常应从第一阶段开始，但很多人并未按规范执行）。 - 深入了解候选人的逻辑推导、在哪些环节犯过错误、遭遇了何种突发问题以及最终如何解决。 - 重点评估“伦理判断”（Ethical Judgment）能力，探讨面对各类 AI 工具时，不仅要看“能不能做”，更要思考“应不应该做”。 ```mermaid flowchart TD accTitle: AI 过度拟合简历与真实评估对比 subgraph S1["过度拟合简历（不良实践）"] A["使用 AI 强行贴合岗位描述"] --> B["表面关键词完美匹配"] B --> C["面试环节暴露虚假技能"] C --> D["丧失真实人类特质与信任"] end subgraph S2["有效评估框架（基于 CPMAI）"] E["六阶段方法论拆解"] --> F["考察真实问题与纠错过程"] F --> G["评估伦理判断与决策逻辑"] end ``` --- ### [33:24](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2004s) 初级员工的技能悖论：工具熟练度与 AI 素养的脱节 现场观众提出疑问：当前新入职的初级员工（Junior Staff）虽然在业务知识和组织架构经验上较为欠缺，但在 AI 工具的实操上可能比资深员工更具优势，企业该如何重新定义“初级”这一概念？ Paul LaForge 针对 Z 世代（Gen Z）的现状做出了风趣而深刻的回应。他指出，年轻一代确实普遍属于“工具达人”（tool savvy），但多项研究表明，他们普遍存在 **AI 素养（AI Literacy）偏低** 的问题，极易盲目信任 AI 生成的输出结果。初级员工往往缺乏质疑精神和底线判断能力（无法敏锐察觉“这个结果不可信、有错误或有缺陷”）。Paul 强调，只有当一个人不仅熟练掌握工具，同时还具备扎实的 AI 素养和批判性审视能力时，才能真正迈向“资深”（Senior）水平。 --- ### [34:50](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2090s) 生产力提升与员工技能流失的矛盾应对 观众 Adam Maine 提出了一个关键悖论：企业引入 AI 虽能提升生产力，但员工原有的专业技能、判断力以及机构沉淀的知识会逐步转移到系统中，企业该如何平衡这种技能流失？ Columbia University 的 Lilian Ajayi Ore 结合安永（Ernst & Young）的最新调查数据指出，许多高管抱怨在 AI 上过度投资却看不到预期回报，单纯强调“缩短流程”或“剔除无关工作”已不再有效。破解这一矛盾的核心在于 **“流程同步”（Synchronize your process）**： 1. **避免单向过度加速：** 不能让技术落地推进得极快，而劳动力技能培训滞后脱节，两者必须协同并进。 2. **统一组织认知与培训：** 就像哥伦比亚大学的做法一样，即使部分员工已是资深玩家（Power Users），也必须与全体团队共同接受培训，因为个人/商业场景下的使用方式与职场和教学场景截然不同。 3. **建立个性化与严谨性反馈：** 好的 AI 应用需要持续喂入明确的行为习惯数据（例如 Lilian 要求自己的 AI 助手必须提供经同行评审的期刊引用与高光标注），在保持步调一致的同时建立严谨的校对习惯。 --- ### [37:15](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2235s) 防范技能退化与拓展蓝领/传统行业的 AI 认知 针对现场观众关于“如何防止因过度依赖 AI 而导致稀缺技能退化（Downskilling）”的提问，专家们给出了具体建议： - **坚持持久技能的日常训练：** 绝不能因为 AI 的输出看起来很合理就放弃批判性思维，必须养成主动质疑、挑战结果真实性的习惯。 - **适度回归传统操作：** 偶尔尝试“不借助 AI 的传统方式”（就像不用计算器徒手算数学一样），体会其中的复杂过程，以此保持底层思维的敏锐度并找到人机协作的平衡点。 - **拓宽对各行各业 AI 应用的宣传：** AI 的影响早已超越科技行业。现代汽车本质上就是一台移动计算机，汽修技师（Mechanics）同样需要接受 AI 相关的技术培训与认证。社会应加大对管道工、技工等各类传统蓝领及实用工种（Trade jobs）在智能化转型中的价值宣传。 --- ### [38:46](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2326s) 专家寄语：职业路径探索与建立核心竞争力 在论坛尾声，主持人 Kristin Stoller 邀请每位嘉宾为正在权衡“学位 vs 证书”的新毕业生及招聘方提供实用建议： - **Aisha Lawrey（约翰斯·霍普金斯大学）：** 很多人的阻碍源于“害怕被淘汰”的恐惧心理。现在的许多岗位在五年前甚至一小时前根本不存在。学生应跳出固有框架，不要局限于申请现有岗位，要勇于探索、开辟属于自己的全新职业赛道。 - **Lilian Ajayi Ore（哥伦比亚大学）：** 强调要在职场中创造双赢（WIN）。AI 基础知识只是门槛，企业真正看重的是员工与生俱来的本能特质、好奇心以及解决实际工作流难题的能力。必须明确技术只是辅助工具，核心竞争力在于如何利用自身技能推动业务改进。 - **Paul LaForge（西部州立大学，WGU）：** 倡导“投资自身专长”的思维模式。求职者需清楚梳理自己能带来哪些不可替代的独特价值。对于初入职场的迷茫期，应积极进行职业探索，不必强求第一步就达到完美，“因为没有人能在第 1 步做到完美，甚至到了第 14 步也未必完美”。 ### [42:05](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2525s) 技能提升与微证书：填补职业鸿沟与自我投资 针对那些已经在既定职业轨道上立足、但希望进一步提升技能（upskill）或实现职业转型的职场人士，专家建议应密切关注相邻的岗位领域，敏锐识别出拉开差距的核心技能（"difference maker"）。虽然在 Western Governors University（WGU）等高等院校中，传统学位依然是教育体系的核心组成部分，但微证书（Certificates）正扮演着日益关键的角色。如果评估后发现个人的技能差距不需要一个完整的学位来填补，而是可以通过专项证书来解决，那么就应当果断选择考取证书，通过这种精准的自我投资来持续增强自身的专业壁垒。 ```mermaid flowchart LR accTitle: 职业提升与技能填补路径 subgraph Gap["现状与差距分析"] A["现有职业路径"] --> B["相邻岗位技能差距"] end subgraph Action["技能弥补方案"] B --> C{"差距类型"} C -->|"深层系统性重构"| D["攻读高校学位"] C -->|"专项技能缺口"| E["考取专业微证书"] end subgraph Outcome["职业发展"] D --> F["增强专业壁垒"] E --> F end ``` ### [42:31](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2551s) 从高中到职场：PMI 项目管理认证与终身学习的“两者皆可”路径 Project Management Institute（PMI）在市场上占据着极为独特的生态位置，能够深度联动高校开展联合培养项目。通过这种合作，学生在本科或研究生阶段攻读项目管理相关学位的同时，由任课教授将认证知识贯穿于整个学术培养周期，使学生毕业时就能直接考取 PMP（Project Management Professional）这一核心职业认证。除了行业标准的 PMP 之外，PMI 还推出了门槛更低的 CAPM（Certified Associate in Project Management，助理项目管理专业人员认证）。 PMI 甚至在部分高中开展了试点项目，只要年满 18 岁（符合参加法律考试的最低年龄要求），学生便可以在没有任何前置经验要求的门槛下直接报考 CAPM。这让年轻人得以在更早的阶段接触并掌握一门专业技能。教育与职业发展可以发生在人生的任何阶段，学历教育与微证书之间绝非零和博弈，而是一种“两者皆可”（"yes and"）的互补关系。无论是重返大学深造还是考取前沿认证，都能帮助求职者探索那些原本无法触及的职业新机遇。 - **PMP 认证（Project Management Professional）**：PMI 的旗舰核心职业认证，通常与高校本科或硕士课程整合，形成“学位 + 认证”的双轨培养。 - **CAPM 认证（Certified Associate in Project Management）**：低门槛入门级项目管理认证，无前置经验限制，年满 18 岁即可报考，目前已在高中国际试点项目中推广。 - **学习范式转变**：职业路径规划应打破“学位 vs 证书”的二元对立思维，将其视为协同增效的终身学习工具箱。 ## Notable quotes > [05:06](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=306s) 绝对不能在没有培训的情况下直接把工具丢给员工 — Kathleen Walch > [05:06](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=306s) AI-first, never start with a blank page — Kathleen Walch > [08:06](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=486s) 先抓 AI 素养建设，然后再跟进工具的引入 — Paul LaForge > [10:58](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=658s) heavy usage that's very shallow — Kathleen Walch > [10:58](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=658s) what's in it for them — Kathleen Walch > [23:08](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1388s) 坦白说，高等教育自身必须被重构。 — Paul LaForge > [23:08](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1388s) 教会一名初级新手，如何以资深工程师的素养与成熟度走出校门。 — Paul LaForge > [28:52](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=1732s) 你个人的独立贡献到底是什么？ — Dr. Lilian Ajayi Ore > [38:46](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2326s) 因为没有人能在第 1 步做到完美，甚至到了第 14 步也未必完美 — Paul LaForge > [42:31](https://www.youtube.com/watch?v=V7WpDqWrZ4w&t=2551s) yes and — Kathleen Walch ## Mentioned Kristin Stoller, Kathleen Walch, Paul LaForge, Dr. Lilian Ajayi Ore, Aisha Lawrey, Adam Maine, Fortune, Project Management Institute (PMI), Western Governors University (WGU), Columbia University SPS, Johns Hopkins University Whiting School of Engineering, Global Connection for Women, Ernst & Young, CPMAI, Microsoft Copilot, Microsoft 365 Copilot, Gemini, PMP, CAPM, Prosci, ADKAR, Quality Matters, AI Today ## Video description Kristin Stoller - Editorial Director, Fortune Kathleen Walch - Director, AI Engagement & Community, Project Management Institute Paul LaForge - Vice President & Dean, School of Technology, Western Governors University Lilian Ajayi Ore - AI Lead Faculty, Columbia University SPS Bestselling Author, The Power of Learning Mindset, Columbia University Aisha Lawrey - Responsible AI Portfolio Lead, Engineering Executive and Professional Education, Johns Hopkins University, Whiting School of Engineering Building an AI-powered company requires more than just buying software; it requires a complete rethink of how talent is deployed. This panel features HR visionaries and operational leaders discussing the "top-down" approach to upskilling. Learn how global organizations are redesigning job descriptions, navigating the transition of traditional roles, and deciding which tasks remain uniquely human. Discover how to turn workforce readiness into a competitive moat and why your "human strategy" is now your most important AI strategy. [Source](https://www.youtube.com/watch?v=V7WpDqWrZ4w)

[5:06] Kathleen Walch: 绝对不能在没有培训的情况下直接把工具丢给员工 [5:06] Kathleen Walch: AI-first, never start with a blank page [8:06] Paul LaForge: 先抓 AI 素养建设，然后再跟进工具的引入 [10:58] Kathleen Walch: heavy usage that's very shallow [10:58] Kathleen Walch: what's in it for them [23:08] Paul LaForge: 坦白说，高等教育自身必须被重构。 [23:08] Paul LaForge: 教会一名初级新手，如何以资深工程师的素养与成熟度走出校门。 [28:52] Dr. Lilian Ajayi Ore: 你个人的独立贡献到底是什么？ [38:46] Paul LaForge: 因为没有人能在第 1 步做到完美，甚至到了第 14 步也未必完美 [42:31] Kathleen Walch: yes and

[0:04] Paul LaForge: AI is so much better if you have a strong data model underneath what your LLM is trying to get at. [23:08] Paul LaForge: we have to teach a junior how to graduate as a senior. [25:22] Lilian Ajayi Ore: If we all hire from the top, we're going to deplete the institutional knowledge that we need in order to thrive and grow our business for years to come. [31:32] Kathleen Walch: If you do that, you're losing your authentic human self, right? Why are we hiring you? Right? Because you're a human. If we didn't want to hire a human, we would create an AI agent. [31:32] Kathleen Walch: just because you can, should you? [33:24] Paul LaForge: I don't trust this, this is wrong, this is bad. [38:46] Aisha Lawrey: Don't be afraid to just step out there and carve your own lane and create your own path and your own career that you could do that may not exist right now. [38:46] Paul LaForge: none of us have and we don't think we're necessarily perfect on step 14 either. [42:30] Kathleen Walch: yes and, not an either/or.

[0:04] Hi everyone. I'm Kristen Stler. I'm the editorial  director at Fortune. Uh, a big part of what I [0:08] focus on is covering the future of work. I write  a Monday newsletter and so this topic is very, [0:14] very near and dear to my heart. Um, I want to get  started with a quick, you know, fire starter for [0:19] all of our panelists. So, we're going to say,  you know, in a minute or less, can you tell me [0:24] your name, your title, your company, and then I  would like to know what do you think is the single [0:29] most overrated skill for the next 5 years and most  underrated skill for the next 5 years. Kathleen, [0:36] let's start with you. All right. Hi, everybody.  I am Kathleen. I am the director of AI engagement [0:42] and community at Project Management Institute.  Today is a very special day for me because it [0:47] marks two years that I've officially been here.  I came from an acquisition with a company that [0:52] I had co-founded. And so with that, we brought a  methodology and certification called CPMAI, which [0:59] is for running and managing AI projects. It's  our fastest growing certification at PMI, which [1:03] I'm really excited about. And also the AI today  podcast, which is on season 10, which is crazy. [1:08] It's a decade. I know. I know. And I have not run  out of things to say. Focused on AI today. So we [1:14] don't really look back into the f or you know back  into the past or too too forwardleaning you know [1:19] thinking about what's going to happen in the next  5 10 20 years it's hard to predict. Uh so that's [1:24] a little bit about myself. I'm also based in the  DC region for anybody that has traveled from the [1:30] other coast. And what is something that I find the  most underrated skill and then the most overrated [1:36] skill. So underrated skills I think honestly to  understand really fundamentally how to run and [1:43] manage AI projects right it still is the fastest  growing certification but folks don't always have [1:48] a grounding in why data is so important figuring  out what business problem you're trying to start [1:54] to solve always always start with that most people  will jump right into phase four which is model [2:00] development that goes for agents as well and then  I think that the most overrated skill right Now, [2:08] I don't know, maybe I mean like basic prompt  engineering. That makes sense. Thank you, [2:14] Kathleen. Paul. Hi. Yes. I'm Paul Leforge. I'm  a VP and dean at the School of Technology at in [2:20] at Western Governor's University. I oversee  computer science, software engineering, AI, [2:25] uh, and our data programs. I, uh, am resident in  Salt Lake City, so just a little ways from here. [2:32] Uh, as far as the most overrated skill,  I would probably go with being able to [2:37] use an AI tool to generate output. Period. Uh,  underrated skill, it's all about the data. Uh, [2:45] people forget about the data. They neglect the  data. AI is so much better if you have a strong [2:50] data model underneath what your LLM is trying  to get at. So, that'd be my answer. Good one, [2:55] Lillian. All right. Good morning, everyone. I am  Dr. Dr. Lillian Aay or I am an AI faculty member [3:02] at Columbia University where I'm also a full-time  lecturer and so wanted to add that I do have a [3:09] nonprofit organization called global connection  for women and we are responsible for providing [3:14] leadership training and even entrepreneurship  training. So when it comes to an underrated [3:19] skill set when we look in the AI space I think  is strategic communication. Um and albeit that [3:26] I teach in a communication program. The reason why  I say street communication is how to create value [3:33] in how we express the need for adoption and how  to express why we need the adoption in the first [3:39] place. And then how do we communicate the value of  getting everybody involved not just specific set [3:45] of our workforces so that we don't lose people in  this AI acceleration that we're facing right now. [3:53] And then the most high value skill I think is the  openness uh which I discuss in my book and I'll [3:59] talk about it later is we are all willing to learn  and we're willing to accept the phenomenon that's [4:05] taking place in our workplaces and I think that  is a skill that we should acknowledge in ourselves [4:10] and give ourselves props for having that level  of mindset as well. Thank you Aisha. Hello good [4:16] morning everyone. My name is Aisha Li. I'm the  responsible AI portfolio lead at John Hopkins [4:23] University in the department of executive um  engineering and professional education and [4:30] um I'm from the DMV um originally from New  Jersey and I'm really happy to be here. I [4:35] would say that you know looking at in this space  of AI something that I see is very underrated is [4:42] really talent strategy. Um all of those skills  around building talent is really underrated and [4:49] overrated is just a technology strategy. Um  people spend too much time I feel in terms of [4:56] you know building infrastructures but not enough  around the talent and the people who are going [5:01] to manage those infrastructures. All right. Thank  you all. Now audience I have a question for you. [5:06] Uh one because we're we're genuinely curious up  on the stage and two because I think we we need [5:10] to wake everyone up. I hope everyone's had their  coffee. But please, I want to see a show of hands. [5:15] Tell me how many of you are investing more in  learning and development this year than you [5:21] did last year. Okay. And how anyone investing  less? Just curious. Okay. No one investing. So [5:30] that that's a good sign for us here on stage, I  think. So I want to start with with a case study [5:35] because I think what really helps people to learn  is hearing about your your trials, your failures, [5:40] but also your successes. Can we start by talking  about you know what is one thing your organization [5:46] tried in the name of upskilling that you thought  was going to work and and really didn't and then [5:51] what worked instead? And whoever wants to start  with that one, we're all going to rat out our [5:58] organizations. [laughter] Um I guess I can get  started on this. So I right I've been here for two [6:08] years now and before that I worked for myself and  so it's been a culture shock for me anybody that's [6:15] gone from working for themselves being their own  boss making all the decisions whether it's right [6:19] or wrong right you feel all the consequences of  that to now working for a larger organization [6:24] whenever people talk about AI change management  usually goes handinhand with that why because [6:29] there's so much change that needs to happen and so  one thing that I did for myself. So, I really like [6:35] that everybody is investing in themselves. I I  am. I think everybody on the panel is as well. So, [6:39] I recently got proi certified because I said  if everybody's talking about change management, [6:43] I want to learn it a little bit better and be able  to really articulate what that means, especially [6:49] in this AI era. So, one thing that we've done,  I'm going to lean with the what we did well. So, [6:56] I always say you can't give people a tool without  training, right? Because now you just give people [7:01] access to something and what happens? they get  busy. they one maybe don't have that awareness [7:06] right if we talk about adcar and two they  definitely don't have the d the desire what's [7:10] in it for them why should they learn why should  they be investing in that particular tool and so [7:15] we are a Microsoft shop so we use co-pilot by show  of hands who here uses co-pilot whether or not you [7:23] want to I have to say it's gotten a lot better in  the past few months like a lot better in the past [7:28] few months uh which I've been really excited  about so I said we can't give people this tool [7:32] unless we have an actual ual roll out and in uh  you know a proper training and so my organization [7:39] said okay and we've invested in that and I have  seen tremendous gains I think people still it's [7:45] I always say it's a muscle right that you have to  reflex be AI first never start with a blank page [7:50] I think a lot of people are there at this point  but still make sure that you use the tools that we [7:55] give you access to and so that was one thing that  I think that we did really well and it could have [7:59] gone very wrong that's a great example thank you  Kathleen Paul, you have one? Yeah. Yeah. At WGU, [8:06] we're heavy on innovation and I think right  now we're at we have at least 76 different AI [8:10] initiatives going throughout the university.  And I confident that's just because somebody [8:14] stopped counting. Uh they knew there was more.  But the problem is in that rush to innovation, [8:19] you can do tool exposure, but AI literacy falls  behind. And so what we're finding ourselves need [8:26] to do is how do we catch up on our AI literacy  to all the innovation that's happening and to [8:31] all the tool usage. And I think if we would have  done it differently and could get have less FOMO, [8:37] we would have led with AI literacy and then  followed it up with the tools. Makes sense. [8:43] Any other examples? Yes, I have some um as  well. And you know when we we were one of the [8:49] later institutions that adopted that allowed our  students to openly use AI because we all knew that [8:55] our students were using it and we you know because  when you're part of a university there are levels [9:01] and chains and commands and things like that.  But when we had permission to have full adoption [9:06] across our university for AI both for the faculty  level and our student level, we did it differently [9:13] with our department because we wanted to confront  what every organization faces with this decide of [9:18] like okay is AI going to do my work better than  me? Is it going to replace my work? And we started [9:24] with AI literacy. So what I did was me and my team  we sat with the IT department that already had the [9:30] AI training and we said yes now the university is  permitting us to use AI in our classroom to allow [9:36] students to use it because one of the reports  that we read said that students feel like faculty [9:41] members antagonizes them and sometimes students  don't use AI and yet because we're all in this [9:46] mindset of AI is helping you do your work that  [snorts] we accuse our students without knowing [9:50] it. So we put together a curated AI literacy  training for our faculty. And so one of the things [9:58] that we had to do was a survey. So you talked  about data earlier. You need to understand the [10:02] readiness of your team of your talent population  whatever skill that they are. And you also need [10:09] to make it seem like an intervention that they  desire and they need. And so I had we had our [10:14] conversation with our faculty say listen this is  not about you know you losing pace because you [10:20] don't know how to use it or you've never used it  in your teaching. This is about giving you more [10:24] tools to be more effective in the classroom. And  so once we broke the barrier and reframed how we [10:30] want AI to be used in the classroom then we offer  the training then we did some data survey to find [10:36] out where the learning gap. And when I talked to  my superior she was like let's make it anonymous. [10:42] So you don't know if it's Lillian that's filling  out the application. So that way there's no [10:46] stigmatization that goes with the development of  this training that we did. And the result was an [10:52] ease in adoption. There was an appreciation for  the adoption. And so we showed them from literacy [10:58] all the way to advance of how to carefully  integrate AI at whatever pace that you're on [11:04] because it's not something that we're forcing on  you. It's something that we want you to see as a [11:07] tool that is capable to help you grow. And I think  the feedback has been very good. We're actually [11:12] going to run the training again and get more, you  know, facilitate. We didn't make it a required um [11:19] training. I think that's something that sometimes  when we're managing t talent in the workforce, [11:24] what is mandatory? What is recommended? It wasn't  mandatory, but it was recommended. And we saw it [11:30] like an additional professional development  tool that our faculty were very receptive [11:34] to and we gave them the language to talk to our  students about it as well. So, it was all success. [11:39] AJ. Yeah, I just think quickly just to piggyback  on even what Paul said, it's that whole, you know, [11:44] AI tools versus the AI literacy. And I think  when um you know, organizations, institutions [11:52] uh spend more time on the technology rather than  the education, you will get resistance, right? [11:58] And then you'll also get overestimation in terms  of the AI capabilities. So we really, you know, [12:05] just kind of took a step back and leaned more into  the education rather than just focusing on the [12:10] technology. I want to talk about that resistance  and and Kathleen, you you kind of touched on this [12:15] too about how do you incentivize people to  do it? Because learning clearly doesn't end [12:20] at the university anymore. You have to keep people  learning their entire life. I spoke to a company a [12:25] few months ago that said, "Okay, what we're going  to do is we're going to tie a bonus toward AI use, [12:31] and if we see that you're using AI, you get a 3%  raise or a 3% bonus because of it." So, how do you [12:38] really incentivize people to keep learning? Is it  bonuses? Is it promotions? Is there something else [12:44] that's that magic thing that you think will work?  You really need to say what's in it for them. [12:51] bonuses get really tricky because when you  measure on AI use what I've seen is you can [12:58] have heavy usage that's very shallow meaning  right so again we're a co-pilot shop so we can [13:06] measure who's using it on what different tools  too right the different applications so who's [13:10] using it in chat who's using it in email who's  using it in powerpoint who's using it in Excel [13:15] a lot of people are using it in email and so  it's going to say that you're daily using it [13:19] right using it to help me rewrite an email. Like,  that is barely scratching the surface. But guess [13:25] what? If that ticks a box that I'm now going to be  measured and get a bonus based on that, then you [13:31] can bet every single day my first email is going  to be regenerated with AI. So, I can tick the box, [13:38] get my bonus, and move on. And you've really  missed why you should use it. So, you have to [13:43] sell what's in it for them. Now, you can do this  depending on how large your organization is. We [13:48] first rolled out right tooling and training  for everybody because I said you can't give [13:52] the tool without the training but now we need  unique like vertical case by case whatever you [13:58] however you want to chop down your organization  so finance is going to have different needs and [14:03] marketing is going to have different needs and  our learning team will have different needs and [14:06] then you provide specific training for them. The  other thing that we've done that's been working [14:10] out really well is we've identified we called them  we originally called them uh co-pilot champions [14:16] but then we moved it to co-pilot power users. Why?  Because we were told well everybody should really [14:20] be a champion of the technology. And I said okay  but it's going to be people that are really diving [14:24] deep into it. So terminology whatever you want  to call it but then they can go out so that if [14:28] you have a specific question because what happens  right you watch a video and these tools change so [14:33] quickly especially Microsoft they love to change  where their buttons are like every single week. [14:37] So what was you know maybe recorded 2 months ago  now suddenly doesn't work and then you get super [14:43] frustrated. So if you have somebody that you can  reach out to and ask maybe have a five minute [14:47] conversation it can save you hours of frustration  and then also very vertical by vertical group [14:54] by group so that you get more of that specific  training that may not be relevant to everybody and [14:59] at the university level I you have to incentivize  both students and faculty. So the three of you how [15:05] are how are you thinking about it? Yeah. Well, it  from the organizational side, we really have taken [15:10] a co-creation model with those who are affected  um or those who are potentially afraid that their [15:17] job might be replaced by AI because they actually  really don't like parts of their job. They would [15:23] love those parts to go away. We had this one  group we're working with, they really hate [15:27] control F. And we've started the anti-control F.  My favorite command, Paul. I I'm very surprised. [15:33] you just do it a little bit too much and it  becomes less you have your favorite and so [15:37] uh but by working by co-creating the strategy  with them bringing them part of that how as [15:42] as we are figuring out a solution that is more  AI le and AI centered uh what are the what are [15:49] the skills that boy I I did know a little bit  of Python but I didn't know I didn't have to [15:52] use it as much when I'm doing control F control  F control F but now that that they're part of [15:57] that process they they're upskilling as we are uh  as we are leading that transformation information [16:03] and what's key and this is what I think a lot of  employers miss out on is your domain expertise [16:09] is sitting right in that chair and if you design  a solution outside of that domain expertise your [16:15] AI transformation will fail and by having those  folks right working with us I think that's been [16:22] key in in terms of the students uh on the student  side it is different uh we're we have uh multiple [16:29] studentfacing AI tools that they can opt in and  choose to use and it's based on right now we we [16:37] offer a lot of voluntary usage. Uh so people I  can test the waters. Oh well that was actually [16:42] kind of nice. All right. And do you get people who  actually voluntarily Yes. Absolutely. Absolutely. [16:47] And they can they like boy I can practice for  something. I have I have somebody who can like [16:50] quiz me now on some things or I'm really stressed  about my finances. Help me out here. And it's 3:00 [16:56] a.m. which is a notorious WGU like working hour.  and can't really call somebody but you can talk to [17:03] somebody. Aisha and Lillian, I I also want to add  on to this question too and say, you know, what [17:08] is what is the metric that you're using to telling  you that that tells you that what you're doing is [17:12] working? What is what is one thing you use to  measure that like yes, our upskilling efforts [17:17] are actually yielding results. So I I think this  so for there if there's any faculty in the room, [17:24] I think for us the true measurement is our course  evaluation. you know, how satisfied are our [17:29] students with the classroom? Do we do they feel  like we're preparing them for the workplace? Um, [17:34] recently I had a coaching session with one of our  AI professors and we had meetings with our board [17:40] members. We went to their offices. We sat with  them and we asked a very simple question. What are [17:45] you hiring for? And we were somewhat surprised  but not too surprised by their responses. [17:51] And one of the things that came up with them is  they're not particularly interested in the school, [17:55] the tools that students are learning in  school. What they need are problem solvers, [18:00] critical thinkers, people who can articulate  how do we solve a problem using all the tools, [18:05] whether it's the AI tools or the traditional tools  that we have in the workplaces. And so, you know, [18:11] our course evaluations are immediate feedback. We  have roundt discussions. We have pedagogical lab [18:17] that provides faculty members with the open forum  to come and talk about whatever challenges they're [18:22] facing. But to add more to the earlier question  about you know how are we thinking about where you [18:29] know what part like how do we incentivize this?  One thing I would say with incentives and this is [18:34] our current leadership challenge right now for us  in our academic institution is if you provide too [18:39] much incentive you create a culture where people  are incentivized and then it becomes what she said [18:44] earlier a check in the box. What you want to do is  create a learning culture that the incentives is [18:50] what you get the output of your effort. But you  do need to recognize give special badges to the [18:56] individuals who are early adopters or champions  ambassadors. But if you create too much incentive, [19:02] you derail the purpose of this tool and the  effectiveness of learning culture to begin with. [19:08] And I'll just add as well, you know, we really  focused on looking beyond just completion rates, [19:13] but also making sure because people feel,  you know, like maybe their roles could become [19:18] obsolete and they're fearful. Um, letting them  have the ability to do more exploring in terms of [19:25] crossf functional collaboration, right? I think  when you give people the opportunity to look at [19:30] redesigning and reaching across you know at the  university maybe in the specific area if they're [19:37] in engineering reaching out to education and  doing more of that crossf functional collaboration [19:43] that's an incentive because you know at times you  know different departments really stayed within [19:48] their department and worked in silos. So that's  another way that we've incentivized not only just [19:53] students even looking for their adviserss their  research but also faculty as well as given them [20:00] more freedom to say you know I want to work with  the school of design or you know being able to do [20:06] more of that collaboration I think that's really  important and we can learn a lot from that as [20:11] well. Absolutely. And we'll get back to the fear  point in a second, but first I want to talk about [20:15] a lot of the buzzwords of skills I'm hearing  right now. You know, critical thinkers, problem [20:19] solvers. Kathleen, you have broken skills into  perishable, durable, and enduring. Correct. So, [20:25] I I want you to talk about that. Where do critical  thinking, judgment, all these very human skills [20:31] fall? And then what do you think companies are  missing right now when they're thinking about what [20:37] skills they need from future generations? Yeah. So  when I think about right AI and lifelong learning, [20:43] what does that look like? And so I've broken it  down into three different types of skills that [20:47] we need. So one is perishable. These are going to  be quick to learn and quick to decay. So what does [20:53] that mean? Right? Keeping up with all of the news,  the latest announcements that come out almost [20:58] daily from all of these different, you know, AI  providers, these AI tools, LLM, Frontier models, [21:05] all of these different things are coming out. Yes,  you need to learn how to use them, but you don't [21:09] need to tool chase. So, if you're a Microsoft  shop, learn C-pilot. If you're a Google shop, [21:15] use Gemini, right? And really learn that before  you go out and you can learn a few of them, [21:20] but if you're going to be learning all the  different tools all the time, you're kind of [21:24] going to lose your mind. Maybe you already have  started doing that, right? And so, I think about [21:28] um that with perishable skills. Then, we have  durable skills. So, these are longer to learn, [21:32] but then have a longer shelf life as well. So  these are around frameworks around methodologies [21:37] like CPMAI methodology which is for running  and managing AI projects. So it teaches you [21:42] that step-by-step approach start with figuring  out what problem you're trying to solve then [21:47] get to your data needs then go to either building  your model or building your agent then testing it [21:52] and then putting it out in the real world.  And lastly we have these enduring skills. [21:56] So these are around those soft skills, right, of  critical thinking, collaboration, communication, [22:02] discernment, judgment. These take longer to  learn, but they're much more transferable [22:08] to whatever role that you're trying to do. And  in order to really be balanced in this AI era, [22:14] you have to have all three. So I think about  it like a dinner plate, right? And if I go to [22:17] a buffet, my children will probably jump right  to the dessert table, love it, but then those [22:23] are going to be those perishables quills, right?  They're going to eat all of the chocolate cake and [22:28] the brownies and the cake pops, all that delicious  stuff, and then they're going to have a crash. So, [22:34] that's you're constantly going to have to  be chasing for your next meal. But, if you [22:38] balance your plate and you have a variety of all  different types of skills, you're going to be [22:43] much more prepared. And I think that goes exactly  to what you were saying where organizations are [22:47] looking for those dur those enduring skills. Why?  Because if you have good critical thinking skills, [22:53] if you have good discernment skills, you have good  judgment skills, you can learn the tools and they [22:58] are going to change and so you're going to have  to keep up with that. But you have those skills [23:02] that help you get to your end point. Thank you,  Kathleen. Now, I'm going to turn to the audience [23:08] um for questions in in one second. So, if you have  one, please raise your hand and someone will bring [23:13] a microphone over to you. But before we do that,  I want to turn to our our three academics here [23:17] because we've all seen the headlines being like  Jenzie and Gen A are it's they they're in freef [23:23] fall. It's going to be terrible for them. They  won't be able to get a job. The entry- level [23:26] career is dead. But then you see actually just  kidding. They'll be able to get a job. It'll just [23:31] look different. So So from where the three of you  sit, talk to me about what you are actually seeing [23:36] on the ground with your early care people who  are launching their early careers. Should they [23:41] be terrified? And you know, how are you preparing  them for this really uncertain narrative that I [23:47] think changes about every week? Yeah, it probably  changed since we started. I think so. Uh yeah, [23:52] there's a I think that I think the big thing that  that we've embraced is that it's not as simple as [23:58] it all sounds. It's not a black and white. There's  multiple causes. There's multiple things happening [24:02] at once. And to really understand how to navigate  it, you need to understand the different threads. [24:06] And so we really are trying to understand those  threads. uh one for instance is just business [24:11] uncertainty uh and that can be because of AI  it could be because of economic conditions and [24:17] so when businesses get uncertain generally the  they don't hire the juniors they back they want [24:23] the seniors and so that can that's partly that's  part of the case and so um but part of that is [24:29] our employer outreach and we have a very strong  employer partnership and helping them transition [24:35] through this uncertain time and helping them  with our AI engineering programs and things [24:39] like that to support their upskilling efforts  to minimize that. Uh, another part of that is [24:46] frankly higher ed has to change. We have to start  expecting more of an apprenticeship or hands-on [24:53] model. And so we're we're pushing ourselves  hard to incorporate experiential learning and [24:58] recognizing almost apprenticeship based learning.  And how can we rethink higher ed because one of [25:04] the more successful ways that the a junior now can  be uh become an a senior overnight is by a mentor, [25:12] someone who is guiding them alongside AI uh to  develop those skills. And of course we have to [25:18] change what we teach. I mean we have to teach  a junior how to graduate as a senior. Yeah, [25:22] that makes a lot of sense. Yeah. No, it's a it's  such a great question and for us, you know, we are [25:29] we understand the investment that our students are  making financially, psychologically, emotionally [25:33] as well when they go and follow their education. I  teach in a mast's program and my program is called [25:40] the school of professional studies. And so one  thing that our dean has always instilled in all [25:46] of us as faculty is that because our students are  coming here to a school that is training them, [25:51] preparing them for the professional world, we  need to be fully equipped to make sure that we [25:57] transfer that knowledge and make sure that in that  six month, four to six month window, our students [26:03] have the competitive advantage to get the job and  not wait a year to land a job as well. So it goes [26:09] back to what you said. One of the things that  I've noticed about my academic institutions and [26:13] we've I've taught in other institutions as well is  everybody is invested in this process. We want our [26:20] talent to be successful outside of our program.  We want them to have the best-in-class education. [26:26] We also included this new model called the quality  matters model that is ensuring us to look at our [26:31] academics whether it's the pedagogy the classroom  environment even the job preparation the alignment [26:37] with the industry to make sure that when our  students leave whether it's in 12-month period [26:41] or a 16-month period they are standing out and  the only way this works is if everybody else is [26:48] invested in the process and I think that mindset  of like we're in this together we're trying to [26:53] create this win model which I explained explained  in my book, we're willing to teach our students [26:58] what they need to have that competitive advantage.  We understand the intentionality that goes into [27:02] building out a pedagogy that creates value for  the industry when they're bringing in new talent. [27:08] We ask the industry to please continue to open up  the entry level positions, the midlevel maticians [27:15] because our students come with different levels of  skills. Some of them have two years of experience [27:19] before they go after their masters. Some of  them are coming directly from college because [27:23] the workplace and the marketplace employment, you  know, market is much difficult than usual. So, [27:30] we're doing our work to make sure that our talents  are top. We're giving them what they need so that [27:35] industry sees the appeal, but we're asking the  industry to keep those doors open for entry [27:40] level so that we continue to build institutional  knowledge as well. If we all hire from the top, [27:46] we're going to deplete the institutional  knowledge that we need in order to thrive [27:50] and grow our business for years to come. I'll  just add to that um what has already been said [27:55] is really focus, we've been really focusing on,  you know, just strengthening those academic and [28:01] industry collaborations, right? Um getting more uh  professionals to come into the classroom, do more [28:08] of the professors of practice for our students  as well as our professionals. We also look at, [28:14] you know, finding what are the best certifications  because that is like really hot for a lot of [28:19] companies right now is looking at, you know, not  only just what are your academic credentials, [28:24] but what certifications do you have to add  to that as well. Do you have any, you know, [28:29] micro credentiing? Um, and those are things  that, you know, we also help to seek out for [28:35] students and our professionals is, you know, what  are those certifications and credentials that are [28:40] really going to help put them in places where  they need to be and give them the ability to [28:44] have more options with more flexibility. So,  those are the things that we focus on. Yeah, [28:49] I love the certifications point. That's a a great  one that I want to come back to. But first, anyone [28:53] out here? I see a qu couple questions in the back.  Hi, thank you for the discussion. For anyone who [28:59] wants to answer, I'm wondering what kind of  questions do you ask candidates that get them [29:03] talking in a way where you can assess their level  of AI fluency and AI literacy? Thank you. Great [29:11] question. Anyone want to take a stab at it? Oh, a  wonderful way is to identify have them think out [29:17] loud for you. And so if they whether it's writing  code, writing writing a prompt and but to say this [29:24] is what I'm thinking right now. this is what I'm  talking about and or this is what this is the kind [29:29] of discussions these are the concerns I see with  the output that I have uh for if I'm again in in [29:35] this is a lean tech here so if I'm looking at code  I architecturally what's wrong here uh using old [29:42] school what are the bad smells in the code make  it better and identify oh I see this problem here [29:48] I see this parameter things like that so I would  say it have the candidate talk out loud uh where [29:54] you can and hear uh not only the their critical  thinking but the why behind the critical thinking [30:02] which gets to judgment things like that answer did  you want to say something yes I wanted to add it's [30:08] a great question and when I had my discussion with  the industry this was something that came up so I [30:13] asked them I said how what's your process and so  right now as you know when you apply for jobs most [30:20] people are getting AI to work on their resume  which is there's nothing wrong with that and [30:24] So employers are actively probing to see can this  person even do the work with or without AI assist. [30:32] And so one of the things and I asked it how  what's your case study for how you recruit what [30:36] questions are you asking do an interview and their  response was like we asked what challenge does [30:41] your business face and how did you implement an AI  strategy to solve the problem what tools whether [30:48] it's proprietary or not how did you what was  your individual contribution and people who are [30:53] passionate about those work know how to express  it which I said the undervalue thin communication [30:58] how can you communicate the value that you're able  to create in your organization, past or present, [31:03] to that employer who's considering you for the  role and maybe ask them to give you two examples. [31:09] So that way you can check for consistency in their  stories as well. So for me, I look at this in two [31:15] different ways, right? How do I use AI tools  to help me be more efficient, do my job better, [31:20] do tasks better, and then also how do I run and  manage AI projects? Right? At Project Management [31:25] Institute, we have to care about projects. And  so I will I will say one using AI to write your [31:32] resume. Has anybody here ever gotten a resume  that seems so when with AI right we have this [31:38] idea of overfitting and underfitting it's been  so overfit to the role that you're like you don't [31:43] have these skill sets and that happened to me and  it's annoying and guess what if you do it you're [31:48] probably not getting calls for interviews.  Why? I had to flag this with my HR team. [31:54] There were two people that they passed through for  an interview with me and I'm like, "These people [31:59] are not qualified." And I don't even know what  questions you asked, but looking at their resume, [32:02] I'm like, I can tell for sure that they have  tailored this so tightly to the job that I'm [32:08] looking for that they're not qualified. So, I  would say that that's a flag. And if you do that, [32:13] you're losing your authentic human self, right?  Why are we hiring you? Right? Because you're a [32:18] human. If we didn't want to hire a human, we would  create an AI agent. We had a keynote about that [32:23] earlier today. Make sure that you're showing your  authentic human self and why you're different. And [32:27] then for me in particular, I ask them to walk  through the methodology, the CPMAI methodology, [32:32] and say, "How have you used this for running and  managing an AI project?" And then that goes back [32:37] to that thought process, too. Right? So, where  did you start? At what phase, right? There's [32:42] only six phases. Start at phase one, but sometimes  people don't. And so, that's how they walk through [32:48] that. and then I can understand their thought  process, understand maybe where they went wrong, [32:53] problems that they ran into and then how they  addressed it. I think one thing um as well is [32:58] just around ethical judgment. Um really focusing  on that because you know the question is for all [33:05] of these different AI tools um just because you  can should you right and so I like to hear you [33:12] know when people talk about projects and you  know maybe just experiences they have to speak [33:16] a little bit around that as well. Great question.  I see another one over here. Quick question on the [33:24] uh uh point about the junior staff. Uh is  it possible that uh you guys are seeing that [33:31] some of the junior staff coming out with a little  bit of AI training or capability might be junior [33:37] in terms of subject matter and institution or  organizational knowledge but might be more senior [33:43] on the AI knowledge, AI tooling, etc. If you could  kind of like ramble on about that a little bit, [33:49] that might be good because we're not quite  sure what that junior term means now. And Paul, [33:54] I think that was your point. Oh, yeah. I was just  ready to jump. Yeah. [laughter] Uh this is uh so [34:00] I guess this is to the Gen Zers. Uh do we have  any in the room by the way? Are those watching [34:05] from home? Okay. Oh, we got one. [laughter] They  to to your point, they generally are tools savvy. [34:13] uh what studies have found is they tend to  be uh very low in AI literacy and too prone [34:21] to trust the output that AI has given them. And  so that gets to uh the enduring skills uh that [34:28] Kathleen was talking about like they they need  to be able to ex to have that sound like this [34:32] I don't trust this this is wrong this is bad uh  and that's they tend to lack that based on some [34:39] research studies. So I think junior would be you  know the tool and you're AI literate at the same [34:45] time and then then you're truly more of a senior.  Yeah, good question. I see we have another one [34:50] over here. Hi, my name is Adam Maine. Um I have a  question. How do you grapple with the paradox of [34:58] adopting AI can increase in productivity while  gradually transferring the employees skills, [35:06] their judgment and their institutional knowledge  into the systems that they use? So, balancing both [35:12] it sounds like anyone want to take a stab at that  one, Lillian? I I'll take a stab at this one. Um, [35:19] one of the things I like about the app for AI4,  good u, you know, endorsement from a faculty is [35:24] that they have a press section and I encourage all  of you to take a look at it because they even have [35:29] research data in there. And so there was a recent  um, survey that was done with Ernest and Young [35:34] and they looked at leaders who are complaining  about overinvestment and not necessarily seeing [35:40] the value that AI allegedly is supposed to bring.  And it's no longer an argument of it, you know, [35:45] it shortens the process. It helps us speed up our  process. It takes all the stuff that's irrelevant [35:50] out. And so I think what needs to happen, and  this is something I'm slowly studying myself, [35:56] is you have to synchronize your process. Um, you  can't have adoption be an over acceleration and [36:04] then the development, which I know you mentioned  the workforce development, be something that comes [36:08] after. It has to be synchronized. And to be  honest, getting your work like similar to the [36:13] example I gave earlier about Colombia and how we  did the survey and then we had the conversation [36:17] then we provided the training and people got  to choose what track they were on. We knew that [36:21] they were power users already in the room but we  forced them to go through the training together [36:26] because the way they use it in the work place in  the classroom is going to be different from how [36:31] they use it personally and in their business. And  so I think the biggest takeaway for you is how are [36:37] you synchronizing how you develop your team, how  you're introducing tools, and how are you creating [36:43] value for both your staff and the institution as  well as the technology itself. Right? I think the [36:50] best AI user would say that the more data it  collects about your habits, the better. My, [36:55] you know, my AI tool knows that I'm I'm all about  the facts. Like don't give me anything without [37:00] any highlights. Do not not include some kind of  peerreview journal. it just knows. And so, how do [37:05] you create that kind of paradox in this space  that requires acceleration is consistency and [37:11] making sure everybody is at the same pace. We love  a fact check. Thank you. Thank you, Lilian. And [37:15] just for the sake of time, I'm going to go to one  more question just because we're coming up on our [37:19] last eight minutes. Hi. We talk about upskilling  but how do you prevent downkilling uh the erosion [37:27] of skills that already short or few supply in the  market and being lost due the use by AI? Well, [37:36] I think it's a habit that every individual  needs to make sure that they're practicing, [37:41] right? This goes back to the enduring skills.  Don't give up critical thinking skills because the [37:47] AI output seems legit. And so I think that that's  really important. And again, practice that right [37:54] and say challenge this. Does this seem right?  And then sometimes too, maybe try and do it the [38:00] old school way. See how painful it is, but it's  not going, you know, then then you can say, okay, [38:05] and you can balance between the two. Like math  without a calculator. I like it. Yeah. The other [38:10] thing that I'll add as well is I think we need to  do a better job of marketing all different types [38:15] of careers, not only just those in technology that  use AI. Um I know like the other day I went to [38:21] take my car in and so the mechanic, the car is a  computer. So it's like everything they still have [38:28] to go back and get training and certification on  AI as well. And so I think we're really missing [38:34] a lot of talking more about how AI impacts, you  know, different trade jobs. We really need that. [38:41] Plumbers, like all of those things. So I think we  need to do a better job of marketing that as well. [38:46] Now, we're very sadly in our last five minutes.  So So with that, I want to give each of you a [38:50] chance to answer in about a minute or less. You've  already given such great advice to our audience, [38:55] but I would like to give another piece of advice  because you and and Aisha, you touched on this. [38:59] You have students struggling with degrees versus  certification. Should I even go to college? What [39:04] is the right path for me? What advice would you  give to and you can choose either these new grads, [39:10] recent grads, or the people looking to hire them  on on where they should be focusing and what's the [39:15] right path? Um Kathleen, we'll start with you and  we'll go down the line. Okay. I was going to say, [39:20] should we start with Aisha first? Oh, we can  we can ending with you. So, let's get the mic [39:25] pressure for you. Go for it. Fine. I would say,  you know, one of the biggest things I I'll I'll [39:31] say is just in terms of like some of the cultural  roadblocks. Um it really is like just fear like [39:37] not fear of the technology, but fear of being  obsolete. And so there are many, you know, [39:43] jobs that exist now that didn't, you know, five  years ago even, you know, like like Paul said, [39:49] I mean, an hour ago, like there are jobs being  created. And so I would say to students, you know, [39:54] really think outside of the box in terms of things  that you can create instead of just applying and [39:59] going for those roles that you know already exist  because technology is moving so fast. So don't be [40:05] afraid to just step out there and carve your own  lane and create your own path and your own career [40:11] that you could do that may not exist right now.  I love it. Thank you, Lillian. Yes. So one of the [40:17] things that I try to instill on all my students  and when I coach faculty I do the same is make [40:23] sure you're creating a win and you know the way  that I describe win is make sure that there's [40:28] an eagerness for people to receive the feedback  you're giving them whether you're recommending the [40:33] use of AI or not and let people discover other  talent and I think you're you're right about [40:38] that. I tell my students all the time that just  because there is an expectation for you to have [40:43] some baseline understanding of AI, it doesn't mean  that your innate skills, these instinctual skills [40:50] that you've developed, even the curiosity that  you have, you have to be willing to express that [40:54] to the organization, they see value in it. And you  know, when I talk to industry folks, I'm sure once [41:00] we all meet up and start talking, you know, in  the workplace, there's lesser conversation about [41:06] technological advancement. We're all talking  about what challenges, what is the workflow, [41:10] how do we improve it, and do we have the right  talent to address these concerns that we have. So, [41:16] I'm very big on identifying what your skills  are and communicating that these things are just [41:21] tools. It enhances your work, but it's not the the  competency that the organization is really looking [41:27] for. Paul, yeah, I would say just really take  the mindset of investing in your own expertise [41:34] and really take stock. What is it that you bring  to the table that no one else does? And if you're [41:39] brand new, it's hard to the workforce because you  you don't know what you don't know. And that's why [41:44] WG we work uh closely with that generation to  be able to help them explore and to be able to [41:50] uh find what might be the best fit. If you don't  have that opportunity, I suggest finding that some [41:54] way uh career exploration and don't feel like  you have to get it perfect on step one because [41:59] none of us have and we don't uh we don't think  we're necessarily perfect on step 14 either. [42:05] uh for those who are more you're you're in an  established career but you want to upskill you [42:09] want to change uh look at those adjacent jobs  and to say what is the difference maker and [42:15] uh as granted degrees are a key part of WGU we  also do certificates but I would say if that [42:19] gap is not a degree but it's a certificate get the  certificate you know what is it fill that gap make [42:25] that investment in yourself so that you enhance  your own expertise yeah and at PMI I think we have [42:31] this really unique area in the market where we do  get to talk to universities and do programs with [42:38] universities as well so that students can come  out with undergrad or graduate degrees in project [42:43] management and have their PMP at the end of it so  that the professor will teach that throughout the [42:49] academic career. We also have an so PMP is our  core right certification for project management [42:55] professionals but we also have CAPM and that has a  lower barrier to entry. So we've actually piloted [43:01] with some high schools as well so that students at  18 can come out sit for their CAPM and then become [43:08] certified. And what is CAPM? CAPM is our certified  associate in project management. So there is no [43:16] prerequisites other than you have to be 18. Why?  Because you have to be 18 to legally sit for the [43:20] exam. And so students are coming out and learning  a profession that maybe they wouldn't have before. [43:26] Right? So education, your professional development  can happen at any stage in your career. Of course, [43:31] I love certifications, right? I think that  it's a great way to continue to upskill as [43:36] universities are evolving and changing. I  always say it's a yes and not an eitheror. [43:41] So for folks that are thinking about, you know,  what is their next step in their career path, [43:45] sometimes you can go back to a university,  other times you can get a certification. Um, [43:49] and it really helps you explore opportunities  you wouldn't have otherwise potentially been [43:53] exposed to. Great advice. Thank you for so  much for being here and sharing all this with [43:58] us. Thank you all for listening. Uh and enjoy  the rest of today's conversations. [applause]