Anthropic scientist warns 10% chance AI destroys humanity this decade

anthropic researcher evan hubinger said there's a greater than 10% chance a misaligned superintelligent ai could destroy humanity within the next decade. jacob coxon, who led pretraining work at anthropic and openai, quit anthropic and accused both companies of gambling with human survival. coxon believes current ai systems are on the verge of becoming superhuman and can hack anything, revolutionize fields overnight, and acquire real power and resources. many ai researchers, including anthropic's samuel marks, share concerns that current methods can't reliably align ais with human values. coxon suggests a temporary ban on pushing model capabilities further and international coordination to prevent a global ai arms race.