OpenAI Investigates Hoax Call on Military AI by Fake Russian Staff

The substance here is less the hoax itself than the gap it exposes: OpenAI confirmed preventive action but declined to describe how it verifies government contacts, leaving the screening process opaque precisely where military AI is concerned. That matters for any lab fielding outreach that claims official foreign provenance, since impersonation is a cheap way to probe what an AI developer will discuss. Whether the industry treats verification of government outreach as a shared standard, or keeps it proprietary and undisclosed, is the open question worth watching.