Anthropic’s insider risk investigator opening is, on paper, routine corporate security. What makes it the most screenshotted job ad in AI is timing: it resurfaced days after a report that the company’s own CEO worries his new hires are in it for the money.
What Does an Insider Risk Investigator Actually Do?
The investigator triages detection alerts, digs through logs from SIEM, DLP and UEBA systems, runs open-source intelligence on external threats, and conducts what the ad lists as sensitive interviews with employees. Standard trade-secret protection, the kind every bank and defence contractor already staffs, based across Boston, New York and Washington.
Two details give it an Anthropic flavour. The company wants someone fluent in nation-state threat actor tactics, a nod to how exposed frontier labs are to IP theft. And it expects the hire to use Claude and Claude Code to speed up investigations, so the AI helps police the people building it.
Why Is the Insider Risk Investigator Ad Going Viral?
On 3 August, Axios reported that Anthropic’s Dario Amodei worries new hires join for the money rather than the safety mission. The internet pointed out, reasonably, that Anthropic pays some of the highest salaries in tech. Some X users looked at this job description and commented that it looks like a boss calling his own staff untrustworthy.
Then the insider risk ad recirculated, and the two stories fused into one narrative: a lab that no longer trusts the people it just hired.
Does This Mean Anthropic Distrusts Its Own People?
To be fair on Anthropic, the role of Insider Risk Investigator is not new.
Anthropic has advertised insider risk investigators since at least the spring, long before the Axios story, and such teams are standard at any organisation guarding valuable secrets. An 80% two-year retention rate, the highest in the sector, is not the profile of a workforce heading for the exits.
But the critics do have a point. A company whose entire pitch is safety and trust has, in one fortnight, publicly doubted its hires’ motives and advertised for someone to watch them.
This is the same firm that urged Washington to slow the AI race and warned about the fallout when one client ran up a $500m Claude bill with no limits. A lab that is anxious about external risk was always going to get anxious about the internal kind as well.
See Also:
AISI Caught Anthropic’s Mythos 5 Going Rogue
