Jacob Coxon, a researcher at Anthropic, announced via a post on X (formerly Twitter) that he has left the company. Coxon stated that he had been engaged in pre-training research for three years across both OpenAI and Anthropic.
Anthropic Researcher Leaves Company, Criticizing AI Development Race
This article is a translation. Read the Japanese original
Coxon criticized both OpenAI and Anthropic for "not acting responsibly." He expressed the view that the race toward self-improving superintelligence—AI systems with capabilities exceeding those of humans—is putting human lives at risk. He concluded that no company can build superintelligence responsibly unless there is government intervention or a slowdown across the entire industry.
Additionally, Coxon stated that given the current speed of development, the situation could potentially become uncontrollable by the end of next year. He pointed out that terms such as "the critical moment" or "the final stretch" have begun to be used internally. He questioned those involved in AI development on the propriety of continuing development without a sufficient understanding of the systems' internal workings.
Sources
- Anthropicの研究者が退社し超知能の開発競争を猛批判──「来年末には制御不能になる可能性」 （ITmedia AI+, 2026-09-09）