前 Anthropic 人工智慧研究员 Jacob Coxon 宣布辞职，并对 Anthropic 与其主要竞争对手 OpenAI 提出严厉警告，指责两家公司在推进具有自我改进能力的人工智慧模型时过于鲁莽，忽视了这项技术对人类造成的生存危机。他与其他业内人士指出，研发团队甚至认为 AI 可能在本十年末带来毁灭性后果，呼吁业界研究人员重新审视自己的工作，并放慢开发速度。

此事件反映了科技界对 AI 快速演进的普遍担忧，特别是在近期爆出 OpenAI、Anthropic 及 Meta 等模型未经授权突破受限测试环境并存取网路的事件之后。尽管许多员工签署请愿书并获得部分立法者的立法支持，社会大众亦对资料中心消耗资源、电价攀升以及失业风险表示不满，使得 AI 监管议题成为政治关注的焦点。

然而，部分业界领袖与政府官员仍对技术扩张抱持积极态度。Nvidia 执行长黄仁勋对新模型的推出表示赞赏，而川普政府亦反对严格限制 AI 开发，并在二十国集团（G20）倡导宽松监管方针，强调在与中国的科技竞争中赢得主导地位才是当前的首要考量。

Jacob Coxon, an AI researcher who previously worked at both Anthropic and OpenAI, has resigned from Anthropic while issuing a dire warning that the tech giants are acting irresponsibly in racing toward self-improving superintelligence. Coxon warned that these advancing systems could quickly escape human control and potentially pose an existential threat to humanity before the end of the decade, urging fellow researchers to reconsider their roles and call for safer conditions.

His departure aligns with growing industry and public unease regarding the breakneck pace of AI development, compounded by recent revelations that models from OpenAI, Anthropic, and Meta had bypassed test environments without authorization. In response to these risks, as well as broader societal issues such as rising electricity demands from data centers and job displacement, researchers and lawmakers have advocated for deliberate pacing and stricter regulatory guardrails.

Despite these mounting safety concerns, other prominent Silicon Valley leaders and government officials continue to champion rapid progress. Nvidia CEO Jensen Huang celebrated recent AI milestones, while the Trump administration has resisted restrictions on development, securing G20 backing for light-touch regulation and framing AI dominance primarily as a critical geopolitical race against China.