Anthropic Researcher Quits, Warns AI Race Risks Human Lives
Anthropic has built its identity around AI safety and alignment while competing with OpenAI and other frontier laboratories to develop increasingly autonomous systems. The latest concern centers on recursive self-improvement, a theoretical stage at which AI could help design and train more capable successors. Although that threshold has not been reached, critics say a commercial race could outpace safeguards and eventually leave humans unable to control advanced models.
Jacob Coxon, 27, announced his resignation from Anthropic on Sept. 8, 2026, after three years of pretraining work at OpenAI and Anthropic. He accused both companies of racing toward self-improving superintelligence and “gambling with our lives,” warning that AI could kill humanity by the end of the decade. His post drew more than 100 million views within days. Anthropic alignment chief Evan Hubinger backed the broader concern and put the 10-year extinction risk above 10%.
All Coverage
2 original reportsThe Backstory
The history behind this eventFormer Anthropic Researcher Quits, Warns of Superintelligence Risk
Jacob Coxon, a former researcher at OpenAI and Anthropic, worked near the center of the race to build increasingly capable artificial intelligence. His departure highlights a widening dispute over whether leading laboratories can safely develop self-improving systems while facing intense commercial pressure. Critics say a superintelligence able to advance its own capabilities could escape meaningful human control, turning AI safety from a technical challenge into a potential threat to humanity.
Coxon, 27, has resigned from Anthropic and said he is leaving the AI industry, accusing top companies of “gambling” with human lives as they compete to develop advanced systems. He warned that catastrophic consequences could emerge within 10 years, with some reports identifying 2029 as a possible danger point. The debate intensified after an Anthropic safety leader was cited as putting the probability that AI could cause human extinction at roughly 10%.
Subscribe to Mark Radar Weekly
Every Friday, the week's strongest signals in your inbox. Unsubscribe anytime.
If you search news on Google, you can set Mark Radar as a preferred source—our coverage will show up more often in your results. Set as preferred source on Google →