Osegi
AI·2 min read

Anthropic is hiring a CBRN-E threat investigator for up to $290,000

Anthropic wants someone to hunt people who try to use Claude for weapons, pathogens or toxins. The pay is $230,000 to $290,000, and the case file behind it is odd.

A security analyst reviewing data on multiple screens in a dim office, representing AI misuse investigation work.
Image: ギズモード・ジャパン

Anthropic, the company behind the AI assistant Claude, has a job opening with a title that sounds like a thriller: Technical CBRN-E Threat Investigator. The salary range is $230,000 to $290,000.

CBRN-E stands for chemical, biological, radiological, nuclear and explosives. The job is to spot, investigate and stop people who misuse Anthropic's AI in those areas. The listing says plainly that this means looking into attempts to use AI for weapons, pathogens and toxins.

What does a CBRN-E threat investigator actually do?

Day to day, it means digging through data with SQL and Python and tracing what a bad actor did. It also means writing reports and working with government agencies and scientific groups. Part of the job is predicting how people might misuse AI as it gets smarter.

The company's own example is a strange one. In May, a user asked Claude to help write a grant application for research on the mosquito-borne chikungunya virus. The research aimed to make it spread more easily between people and slip past human immunity. A system that watches for risky biology conversations stopped the request.

The affiliation raised more flags. The application named a military research institution, yet the listed researcher was a civilian. That mismatch started the investigation. It turned up infrastructure in the US used to get around regional restrictions, grey-market resellers and fake accounts. All the related accounts were suspended.

The people involved are working scientists, and the report does not claim they meant harm. Biology can protect or hurt, and telling the two apart is often hard. That is the job.

Pay, location and requirements

The listing asks especially for deep expertise in chemical defence or biodefence. Preferred qualifications include hands-on experience countering weapons of mass destruction and an active Top Secret security clearance. The roles are in San Francisco, New York and Washington DC, with at least 25% of time in the office.

Anthropic is not alone. OpenAI has also advertised for a biological and chemical risk researcher, at $295,000 to $445,000. Applicants are warned they may see explicit content on many topics and may be asked to handle escalations on weekends and holidays. Experts are split on the report, with some calling it a chilling warning and others calling it overblown.

Apps worth a look

SUMO: the number puzzle

Tap the right tiles, add them up, hit the target. Easy to learn, hard to master. Free, for all ages.

Tags:anthropicclaudecbrnai safetyjobsbiosecurity

More from this day

See all →

More in: AI

See all →