Anthropic’s head of safety has stated that there is a greater than 1 in 10 chance that artificial intelligence could lead to the extinction of humanity. The assessment comes amid a period of internal transition at the AI research company, following the recent departure of a colleague.
What Happened
According to reports, Anthropic’s safety leadership has publicly quantified the risk of AI-driven human extinction. The figure cited is a probability of more than 10%. This statement was made shortly after a notable resignation from the company's safety team.
Why It Matters
The estimate reflects a specific concern regarding the alignment of frontier models as articulated by Anthropic's safety leadership. The statement underscores the potential for catastrophic failure modes in advanced systems, as identified by the company's internal risk assessment.
The Bottom Line
Anthropic’s safety lead maintains that the risk of AI causing human extinction is non-negligible, estimating it at over 10%. This perspective, articulated following a colleague's exit, reflects the company's focus on long-term safety and existential risk mitigation in the development of advanced AI.