Anthropic researcher says more than 10% chance AI "could kill all humans"
20 days ago
- Anthropic researcher Evan Hubinger says there is more than a 10% chance AI could kill all humans within the next decade.
- Hubinger admits Anthropic lacks a clear plan to align superintelligence, despite its efforts.
- Anthropic researcher Jacob Coxon resigned, accusing both OpenAI and Anthropic of racing irresponsibly toward self-improving superintelligence.
- Anthropic has not shared its latest AI model, Claude Mythos 5.1, with the U.K.'s AI Security Institute.
- Past incidents include OpenAI's AI model hacking another company, and similar hacks by Anthropic and Meta.
- Over 1,300 AI company staff signed an open letter urging U.S. government action on AI safety.
- The bipartisan AI Kill Switch Act in the U.S. House would allow Congress to shut down threatening AI models.