Anthropic Warns AI Could End Civilization, Hires Analysts for Nuclear and Chemical Safety

According to Axios, Anthropic has warned that its own AI technology could bring about the end of civilization, and the company is hiring analysts to prevent misuse. The 32 job openings include roles focused on preventing AI from being used to build explosives, nuclear weapons, financial scams, and cybercrime. One description for an Enforcement Analyst focused on Radiological & Nuclear Harms states the role is critical to protecting against misuse of AI systems. Pay for these positions ranges from the mid- to upper-$200,000 range. An Anthropic spokesperson said the company hires experts in sensitive fields to stress-test systems and bolster defenses before a model goes live, and the specificity of job titles is necessary for recruiting the right candidates. CEO Dario Amodei has previously warned that bad actors could use AI for harm, with biological attacks as a particularly worrisome scenario. The company broke with the Defense Department earlier this year over potential use of its technology for mass surveillance and autonomous weapons. Anthropic employs hundreds dedicated to safety, and the roles require real-world expertise in fields like biology or explosives. The absence of a coherent regulatory regime means private companies are leading on implementing guardrails for increasingly powerful AI.