The AI Researcher Who Just Quit Anthropic Says It’s ‘Crunch Time for Humanity’

刚刚从 Anthropic 离职的 AI 研究员称：人类正处于“关键时刻”

Artificial intelligence researcher Jacob Coxon sent shock waves through Silicon Valley and beyond on Tuesday by announcing his resignation from Anthropic and delivering a grave warning that the AI race is putting all of our lives at risk. In his post on X, which now has more than 100 million views, Coxon wrote that many of the people building AI share his views and believe time is running out to ensure AI systems are built safely. 周二，人工智能研究员 Jacob Coxon 宣布从 Anthropic 离职，并发出严厉警告称，AI 竞赛正将我们所有人的生命置于危险之中，这一消息在硅谷及更广泛的地区引发了轩然大波。Coxon 在 X 上发布的帖子目前浏览量已超过 1 亿次。他在文中写道，许多从事 AI 开发的人员都持有与他相同的观点，并认为确保 AI 系统安全构建的时间已经所剩无几。

“The consensus is that the next year or two is crunch time for humanity,” Coxon, who worked on the pretraining stage of AI development, said in an interview with WIRED. “These are actually just literal quotes from my colleagues at Anthropic. They’ll say things like ‘endgame’ or ‘crunch time,’” he says. “From their perspective, this is when Anthropic and its competitors decide the fate of humanity.” “目前的共识是，未来一两年是人类的关键时刻（crunch time），”曾从事 AI 开发预训练阶段工作的 Coxon 在接受《连线》（WIRED）采访时表示。“这些实际上就是我 Anthropic 同事们的原话。他们会说‘终局’或‘关键时刻’之类的词，”他说。“在他们看来，这就是 Anthropic 及其竞争对手决定人类命运的时刻。”

It’s far from the first time someone has sounded the alarm about AI, but it comes at a delicate moment. Silicon Valley is scrambling to reckon with the safety and security concerns of advanced AI models. OpenAI has rushed to respond to a security incident in which its agents hacked the platform Hugging Face. Meanwhile, Anthropic is trying to assure investors it has these concerns under control as it reportedly prepares to file for what could be the largest IPO ever. 这远不是第一次有人对 AI 发出警报，但此时正值一个微妙的时刻。硅谷正忙于应对先进 AI 模型带来的安全与保障问题。OpenAI 刚刚紧急处理了一起安全事件，其智能体（agents）入侵了平台 Hugging Face。与此同时，据报道，Anthropic 正在准备进行可能是有史以来规模最大的 IPO，并试图向投资者保证其已将这些担忧控制在可控范围内。

What’s become clear in the response to Coxon’s post is that his views are indeed shared by many of his peers. Evan Hubinger, the AI alignment lead at Anthropic, predicted in a post on X that there’s a greater than 10 percent chance that AI could kill all people in the next decade. That post was reposted by current and former researchers from OpenAI and Anthropic, some of whom said it was a common sentiment in the industry. 在对 Coxon 帖子的回应中，显而易见的是，他的观点确实得到了许多同行的认同。Anthropic 的 AI 对齐负责人 Evan Hubinger 在 X 上发帖预测，未来十年内 AI 导致全人类灭绝的可能性超过 10%。该帖子被 OpenAI 和 Anthropic 的现任及前任研究人员转发，其中一些人表示，这在业内是一种普遍情绪。

What’s less obvious is how exactly these AI fears will come to pass and what the world is supposed to do about the concerns being raised by the people building AI. Coxon, who also worked at OpenAI, tells WIRED that threats could manifest through AI-enabled biological threats or cyberweapons. As a first step, he recommends that OpenAI and Anthropic coordinate on limiting recursive self-improvement—the industry term for when AI is used to build new AI systems. Down the line, he thinks coordination among international power players, including the US and China, will be necessary. 目前尚不明确的是，这些对 AI 的担忧究竟会如何演变，以及世界该如何应对 AI 构建者们提出的这些问题。曾在 OpenAI 工作的 Coxon 对《连线》表示，威胁可能通过 AI 驱动的生物威胁或网络武器表现出来。作为第一步，他建议 OpenAI 和 Anthropic 协调限制“递归自我改进”——这是业内术语，指 AI 被用于构建新的 AI 系统。从长远来看，他认为包括美国和中国在内的国际大国之间进行协调是必要的。

Coxon notes that incidents like the Hugging Face hack factored into his decision to raise alarm bells on the AI race. He also cites the explosive growth of the industry: It now underwrites a meaningful share of US economic growth and has billions of users, while data centers have turned it into a political problem in dozens of states. Coxon 指出，像 Hugging Face 入侵事件这样的事故，是他决定为 AI 竞赛敲响警钟的原因之一。他还提到了该行业的爆炸式增长：它现在支撑了美国经济增长的很大一部分，拥有数十亿用户，而数据中心已使其成为数十个州面临的政治问题。

He claims that, in his experience, Anthropic operates more responsibly than OpenAI, but he expects both companies could cut corners in the future if nothing is done to slow their race for dominance. 他声称，根据他的经验，Anthropic 的运营比 OpenAI 更负责任，但他预计，如果没有任何措施来减缓它们对主导权的争夺，两家公司未来都可能走捷径。

OpenAI and Anthropic did not immediately return WIRED’s request for comment. OpenAI 和 Anthropic 没有立即回复《连线》的置评请求。