
Below is my selection of five recent papers that make meaningful contributions to the current AI governance debate:以下是我精选的五篇近期论文,它们对当前的 AI 治理辩论做出了有意义的贡献:
“AI Wardens,” by Mihailis Diamantis and Sebastian Benthall - link《AI 监管者》(AI Wardens),作者:Mihailis Diamantis 和 Sebastian Benthall - 链接
“AI Agents Push Humans Out of the Loop” by Margaret Mitchell, Avijit Ghosh, and Samir Passi - link《AI 智能体将人类排除在回路之外》(AI Agents Push Humans Out of the Loop),作者:Margaret Mitchell、Avijit Ghosh 和 Samir Passi - 链接
“Human Roles Across the AI Lifecycle: Shifting Beyond the ‘Human Supervision’ Paradigm” by Heni Rothstein Cohen - link《AI 生命周期中的人类角色:超越“人类监督”范式》(Human Roles Across the AI Lifecycle: Shifting Beyond the ‘Human Supervision’ Paradigm),作者:Heni Rothstein Cohen - 链接
“The Autonomy Paradox: Artificial Intelligence and the Foundations of Political Behavior” by Christopher J. Anderson and Elena Pro - link《自主性悖论:人工智能与政治行为的基础》(The Autonomy Paradox: Artificial Intelligence and the Foundations of Political Behavior),作者:Christopher J. Anderson 和 Elena Pro - 链接
“Blameless Responsibility for Faultless AI Harm” by Anna K. Boos - link《无过错 AI 伤害的无责责任》(Blameless Responsibility for Faultless AI Harm),作者:Anna K. Boos - 链接
-
Recommendations:建议:
For everyone: if you find a paper valuable, I highly encourage you to comment publicly on it, share your perspective, tag the authors, and help advance the conversation on AI policy, ethics, regulation, and governance.致所有人:如果您觉得某篇论文有价值,我强烈建议您公开评论、分享您的观点、标记作者,并助力推动关于 AI 政策、伦理、法规和治理的讨论。
For those enrolled in my live training program: this list is a great complement to the material and helps you stay up to date. Bookmark the papers that interest you to read after the program.致参加我直播培训课程的学员:这份书单是对课程材料的极佳补充,有助于您保持与前沿同步。请收藏您感兴趣的论文,以便在课程结束后阅读。
For subscribers attending my live workshops and recorded classes: these are great papers to review over the next few days before the next class launches致参加我直播研讨会和录播课程的订阅用户:这些是非常棒的论文,适合在接下来几天、下一次课程开始前阅读复习。
Happy reading!阅读愉快!
















Thanks Luiza for this list! Looks super interesting, I'll dig in right away :)
Yes, human oversight and legal accountability can be maintained, but only if the AI Agent is not given complete autonomy as an independent decision-maker.