Summary (tl;dr)

AI researcher Jacob Coxon has resigned from Anthropic and left the AI industry, publicly warning that major AI labs are irresponsibly racing towards self-improving superintelligence, potentially "gambling with our lives."

Essential Background

Jacob Coxon, a 27-year-old AI researcher, previously contributed to OpenAI's GPT-4o system card before moving to Anthropic. For three years, he conducted pretraining research at both leading artificial intelligence companies. His work focused on foundational aspects of large language models and their development.

The Full Story

Jacob Coxon announced his resignation from Anthropic on September 8, 2026, and declared he is leaving the AI industry entirely. In a public statement made on the social media platform X, Coxon criticized both OpenAI and Anthropic, asserting that "Neither company is acting responsibly. They are racing straight to self-improving superintelligence and gambling with our lives". He warned that advanced AI systems could become superhuman, capable of hacking, revolutionizing fields, and acquiring significant power, potentially leading to catastrophic consequences for humanity by the end of the decade. Notably, Evan Hubinger, Anthropic's Alignment Science Lead, publicly supported Coxon's concerns, stating his personal belief that "AI could kill all humans!" with a "greater than 10% within the next decade" probability, and that Anthropic currently lacks a clear plan to solve superintelligence alignment. Coxon's resignation comes as Anthropic prepares for an Initial Public Offering (IPO).

Why It Matters

Coxon's high-profile resignation and stark warnings are significant because they come from an insider who has worked at two of the most prominent frontier AI labs. His departure and statements directly challenge the perceived safety narratives of companies like Anthropic, which was founded with a focus on trustworthy and safe AI, potentially impacting its brand value and investor confidence ahead of its IPO. The incident highlights deep-seated concerns within the AI development community about the rapid, competitive race towards increasingly powerful AI systems without adequate safety measures. This internal alarm could intensify regulatory scrutiny on generative AI development and raise public awareness and debate about the existential risks associated with unchecked AI advancement.

Geographic Location

• San Francisco, San Francisco County, California, United States (Jacob Coxon's base as an AI researcher, location of major AI companies like Anthropic and OpenAI)
• Virtual/Online (Announcement of resignation and warnings made on the social media platform X)