Anthropic Exit Sparks AI Safety Debate
AI safety concerns grow as Anthropic resignation spotlights frontier model risks See why researchers warn of superintelligence dangers and a possible AI slowdown
A resignation from Anthropic has reignited debate over AI safety and the risks of frontier model development. Researcher Jacob Coxon said he left the company because he believes Anthropic and OpenAI are racing ahead in ways that could be dangerous.
The discussion escalated after Anthropic’s alignment lead, Evan Hubinger, said he believes there is more than a 10% chance that advanced AI could kill all humans within the next decade. He also said current models are not the main concern, but that selfimprovement and superintelligence remain unresolved risks.
Coxon urged a coordinated slowdown in model capability growth, arguing that avoiding a global race may require temporary limits. The episode has drawn attention because it reflects how seriously some AI researchers are treating longterm safety concerns inside leading labs.