CSET 研究员撰文：AI 灭绝论假设链条脆弱，自动化偏见才是当下真风险
CSET 和 CFR 的两位研究员写了一篇反常识文章：别盯 AI 灭世剧本，自动化偏见这种日常风险正在真实伤人。做 AI 落地、写合规政策的人都该看看，比生存风险辩论更贴近你手头的工作。
CSET 的 Lauren Kahn 与 CFR 的 Michael Horowitz 在新文章中认为，AI 导致人类灭绝的场景依赖一条长假设链，其中每个环节单独看都极不可能发生，同时全部成立的可能性更低。两人并非否认 AI 存在危险，而是认为政策圈过度聚焦低概率、高影响的末日情景，反而忽视了已经在造成伤害的日常性 AI 风险。
他们向华盛顿政策制定者提出的核心关切是自动化偏见（automation bias）：人类对 AI 的信任和任务委托程度超过模型实际准确率，把认知工作整体外包给机器，有时甚至在出现与之矛盾的证据时仍然如此。作者主张监管注意力应转向这一已经发生、可被观察和度量的危害，而不是遥远的生存性风险。
这篇观点与当前美国关于前沿模型生存风险的辩论形成对照，为立法者和机构提供了一种以现有危害为中心的监管切入点。对企业用户而言，自动化偏见直接关联人机协作流程的设计与审计。
An AI-induced human extinction scenario rests on a long chain of assumptions, each highly unlikely on its own and far less likely to align all at once. That’s not to say AI isn’t dangerous. But according to @Lauren_A_Kahn and @mchorowitz (@CFR_org), by fixating on low-likelihood, high-impact scenarios, we miss the forest for the trees: the common, mundane AI dangers already present and actively causing harm.
They urge Washington to focus on automation bias—the phenomenon where humans over-rely on or over-delegate to AI relative to the accuracy of AI models, off-loading cognition to the machine, sometimes even in the face of contradictory evidence. https://t.co/i9wGVe7Yvq
This report is categorized under AITOP's "Papers" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI regulation, compliance frameworks, and sovereign compute policy and US-China foundation model benchmarks and commercialization comparison. Key entities and technical topics under continuous tracking include CSET, 自动化偏见, AI安全, 政策监管.
Source attribution: curated from CSET Georgetown (@CSETGeorgetown).
