Anthropic Safety Lead Estimates Over 10% Chance AI Could Kill All Humans
A senior safety researcher at Anthropic, Evan Hubinger, has stated there's more than a 10 percent chance artificial intelligence could kill all humans by the end of this decade. This comes after Jacob Coxon, another Anthropic researcher and former OpenAI employee, resigned over concerns that AI companies are recklessly racing to develop self-improving systems they cannot control. Coxon accused Anthropic and its rivals like OpenAI of gambling with human lives in their pursuit of advanced AI models. Hubinger acknowledged the risks but admitted Anthropic does not yet have a clear plan for ensuring safety or alignment with human values, despite preparing for potential IPOs. The incident highlights growing concerns within the industry about the rapid development of increasingly sophisticated AI systems and recent issues involving rogue agents. The situation underscores mounting pressure on companies to address AI safety as they push forward in what Coxon described as a dangerous race toward self-improving superintelligence.
Topics: Artificial Intelligence Company Safety Practices Ethics in Tech