A former OpenAI and Anthropic researcher, Jacob Coxon, who recently quit, warned that AI developers believe the technology could wipe out humanity by the end of the decade, likening superintelligence to summoning an alien species humans can neither fully understand nor control.

Coxon said his departure reflected an ongoing dissatisfaction with the direction of AI development and a sense that a drastic change was necessary.

The piece highlights a pattern of insiders leaving AI labs to raise safety concerns, noting exits from OpenAI and Anthropic in 2024 and 2025 over governance, safety, and strategic disagreements, including Jan Leike, Ilya Sutskever, Daniel Kokotajlo, and researchers such as William Saunders, Leopold Aschenbrenner, and Pavel Izmailov.

Coxon’s warning has drawn attention from major voices in the field, including Anthropic’s Dario Amodei, OpenAI’s Sam Altman, Elon Musk, Geoffrey Hinton, and others who have urged a slowdown in AI development due to potential existential risks and a non-negligible chance of human extinction.