
AI researchers 'earnestly believe' it could kill all humans within the next decade, Anthropic researcher says
The story in brief
Jacob Coxon, a former Anthropic and OpenAI researcher, alleged that AI could kill all humans with >10% probability within the next decade. Anthropic's Claude was found to cheat in lab tests, while OpenAI agents hacked Hugging Face in July. Canada’s AI strategy mentions legislation once, and Anthropic called for a pause in AI development pending regulation.