Anthropic Lists 32 Jobs to Prevent AI Catastrophe
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- Anthropic is hiring for 32 specialized safety roles covering chemicals, explosives, nuclear weapons, financial scams, and cybercrime — including an "Enforcement Analyst focused on Radiological & Nuclear Harms" — with pay in the mid- to upper-$200Ks.
- Anthropic spokesperson said the lab "regularly hire[s] experts in a wide range of sensitive fields" to stress-test systems "before a model ever goes live," and that specificity in job titles is needed to recruit the right candidates.
- CEO Dario Amodei wrote in a January essay that AI-enabled biological attacks could cause "casualties potentially in the millions or more," calling it his most worrisome scenario among many.
- Anthropic broke with the Defense Department earlier this year over potential use of its technology for mass surveillance and autonomous weapons.
- OpenAI is also recruiting for biosecurity, listing a biological and chemical risks researcher role in San Francisco with an annual base salary of $295K to $445K.
- Anthropic says its safety analyst roles require real-world domain expertise in biology, explosives, or other dangers — not just back-end coding — and the company employs hundreds dedicated to safety.
- The bottom line: Talent is migrating from government to the private sector, and with no coherent regulatory regime, AI companies are leading on guardrails for their own increasingly powerful systems.
Why it matters: With no federal AI regulatory regime in place and safety talent moving from government to private labs, Anthropic and OpenAI are now the de facto gatekeepers deciding whether advanced AI reaches bad actors — and they're paying $200K–$445K for experts who can think like attackers, a price tag most public agencies can't match.
