AI researchers quit over fears the industry could destroy humanity
Jacob Coxon quit Anthropic this week, claiming AI labs are gambling with human survival.
Researcher Jacob Coxon resigned from Anthropic this week, accusing the company and rivals of gambling with human lives. Current Anthropic researcher Evan Hubinger said the firm truly believes AI could kill all humans.
Multiple other researchers have already quit AI labs over extinction concerns: safety chief Mrinank Sharma left Anthropic warning the world is in peril, while others departed OpenAI and Google DeepMind over similar fears.
AI labs keep building anyway. Anthropic's Dario Amodei estimates a 10 to 25 per cent chance of things going really badly.
Leaders cite three reasons for pushing ahead: they believe the benefits outweigh the risks, safety can only be tested by building AI itself, and they fear losing a race to build powerful AI first.
- 10 to 25 percent chance
- Extinction risk estimate
- Dario Amodei, Evan Hubinger
- Anthropic leaders
Why it mattersIf the engineers building AI think it poses an extinction risk, their decision to continue carries stakes for all of humanity.
