AI Researcher Quits, Says It Could Become Superhuman

Jacob Coxon, a researcher who spent three years working on AI pretraining at Anthropic and OpenAI, has quit his job.

Coxon warns that both companies are racing to develop self-improving artificial intelligence. He says the race could create serious risks for humanity.

He believes advanced AI systems could soon surpass human intelligence. They could also break into computer systems, accelerate scientific discoveries and gain access to real-world power and resources.

Coxon says some employees at leading AI companies already fear that advanced AI could threaten humanity before 2030.

He also argues that many OpenAI employees do not fully understand the potential dangers. In his view, Anthropic has a better understanding of AI risks but feels trapped in the competition.

Coxon says Anthropic may fear slowing down because competitors would continue developing more powerful AI.

He believes the AI race is too important to leave entirely to private companies. He suggests governments may need to temporarily slow or restrict the development of advanced AI.

Coxon is also urging AI researchers to consider whether they want to help develop superintelligent AI under the current conditions.

His warning highlights growing concerns about AI safety, uncontrolled AI development and the future risks of artificial intelligence. .. CN report, 09 Sep 2026