No Safety Plan for AI
An insider just gave AI a 10% chance of ending humanity—and admits there is no safety plan. Here is what is happening behind closed doors.
What the video says
An Anthropic safety leader just gave AI a 10% chance of killing us. This warning from Evan Hubinger followed the resignation of researcher Jacob Coxon.
Coxon accused Anthropic and OpenAI of racing towards superintelligence and gambling with human lives. Hubinger agreed with the threat, admitting Anthropic has no plan to keep advanced AI safe and is not on track to make one.
