Nvidia wants the industry to agree on how to stop AI agents from going rogue. OpenAI didn't sign up in public.
Nvidia has rolled out an industry-wide push called the Open Agent Safety Platform, aimed at curbing unpredictable AI agents. OpenAI is not listed among its public supporters. But according to TechCrunch, OpenAI is privately working with Nvidia on the effort anyway. The report does not specify what that collaboration involves, which other companies have joined publicly, or what the platform requires of participants.
Why OpenAI would work behind the scenes instead of co-signing in public isn't confirmed by the reporting, so treat this as our read rather than a stated fact: aligning publicly with a rival's branded safety framework could hand Nvidia a marketing win that OpenAI may want to avoid, even while cooperating on the technical work itself. It's also possible OpenAI simply prefers to shape agent-safety norms through its own channels rather than someone else's platform.
Either way, the split between quiet cooperation and public silence is the more interesting story here than the safety platform itself.