Anthropic Ex-Employee Warns AI Developers Believe Human Extinction Is Likely Within a Decade

A former Anthropic employee raised alarms on X about the existential risks posed by artificial intelligence. The post highlights a growing concern among developers that AI systems could cause human extinction within the next decade. This perspective shifts the conversation from technical capabilities to immediate safety implications for society.

Internal fears at Anthropic reflect broader industry calls for regulation

Anthropic, the company behind the Claude language model, is at the center of this debate. The former employee noted that developers at the company seriously consider or even believe extinction is a likely outcome. This internal view reflects a broader industry trend where major AI firms are slowing development and calling for government regulation.

Different AI models provide varying estimates of this risk. ChatGPT estimated a 5% to 10% probability of human extinction within ten years. Gemini provided a higher estimate of 10% to 20% based on expert consensus. Claude confirmed the risk is real and non-negligible but did not assign a specific percentage.

These probability estimates vary significantly between models and are based on training data rather than empirical fact. The exact likelihood of extinction remains uncertain, but the consensus is that the threat is tangible. AI development has created a scenario where safety concerns are now central to industry discussions.