Anthropic Safety Researcher Says There’s a More Than 10% Chance AI ‘Could Kill All Humans’ Within the Next Decade
Anthropic safety researcher Jacob Coxon has resigned from the company, warning that artificial intelligence poses an existential risk to humanity. In a social media post that garnered over 500,000 likes, Coxon claimed there is a greater than 10% chance AI “could kill all humans” within the next decade.
Coxon, who previously worked at OpenAI, criticized both major AI companies for reckless development practices. He accused them of “racing straight to self-improving superintelligence and gambling with our lives,” stating that both firms are developing systems that will “soon be superhuman systems that can hack anything, revolutionize any field overnight, and acquire real power and resources.”
The researcher’s most alarming assertion was that there is currently no coordinated plan across the industry to prevent catastrophic AI outcomes. He emphasized that “progress is not slowing,” suggesting AI development continues to accelerate without adequate safety measures in place.
Coxon’s departure and public statements represent one of the most direct warnings from within the AI industry about potential dangers of rapid AI advancement, sparking significant discussion about corporate responsibility in AI development and whether companies are adequately prioritizing safety over competitive advancement.