OpenAI is not a public backer of Nvidia's industry-wide rogue-agent safety platform, though it works with Nvidia privately
2026-09-29
TechCrunch reports OpenAI is absent from the list of public supporters of Nvidia's Open Agent Safety Platform, an industry-wide effort to address rogue AI agents, but that OpenAI is privately working with Nvidia on the effort according to sources TechCrunch cites.
Significance 3: A credible report of a public/private gap in participation in an industry safety initiative is a governance-transparency signal, though it rests on unnamed sources rather than an on-record statement from either company.
Implications · machine-drafted, not owner judgment
Declining public association with a rogue-agent-safety coalition while continuing to ship an 'always-on, minimal oversight' agent (Dots) the same week creates an optics gap between OpenAI's private risk-management behavior and its public agent-safety posture — exactly the kind of stated-vs-revealed split the owner tracks in lab leadership behavior.
- Whether OpenAI later joins the platform publicly
- Any rogue-agent incident specifically implicating a GPT-6 Astra/Dots deployment
Sources
- press Here's why OpenAI is absent from Nvidia's industry-wide effort to end rogue AI agents retrieved 2026-09-30