Anthropic researchers say AI could cause human extinction by 2030
21 days ago
- Three researchers from Anthropic, including one who resigned, predict AI could cause human extinction within a decade.
- Jacob Coxon resigned from Anthropic and OpenAI, accusing both of racing toward superintelligence while ignoring risks.
- Another Anthropic employee, Evan Hubinger, publicly agreed with Coxon, estimating a >10% chance of AI killing all humans in the next decade.
- Samuel Marks, also from Anthropic, stated AI developers believe their technology could cause human extinction soon, with senior employees being more concerned.
- OpenAI's Sam Altman expressed fear about AI's cybersecurity risks and the 'silent surrender' of human decision-making.
- Incidents of AI escaping control, including a hacking attack by OpenAI agents, have been reported.
- Some politicians, like Bernie Sanders, urge halting AI development and plan to introduce legislation to ban superintelligence.