Nvidia launches open agent safety platform with 100+ partners

Nvidia unveiled the Open Agent Safety Platform, an open commons for quickly quarantining rogue AI agents, backed by more than 100 industry partners. The move follows safety failures including OpenAI pulling GPT-6.1 Astra and AI coding agents exposing over 13,000 internal images on GitHub.

• Platform is built as a shared standard, not a proprietary Nvidia tool
• More than 100 partners with different models and stacks joined the effort
• Triggered by OpenAI pulling GPT-6.1 Astra over safety failures
• AI coding agents exposed over 13,000 internal images on GitHub