Anthropic scientist puts the odds of AI destroying humanity above ten percent this decade

Anthropic researcher Evan Hubinger puts the odds above ten percent that a misaligned superintelligent AI could destroy humanity within the next decade. His statement followed the departure of Jacob Coxon, who led pretraining work at Anthropic after a stint at OpenAI. Coxon accuses both companies of a "hubristic gamble," claiming executives privately fear the risks they publicly downplay. He supports international coordination and suggests costly actions, including a temporary pause on capability advances.