As frontier AI labs disclose more agent hacking incidents, Nvidia has launched a new software platform for containing agents with the backing of a group including Anthropic, Microsoft, and SpaceX, but not OpenAI.
The chip company said the Open Agent Safety Platform combines the open source OpenShell secure runtime for AI agents and Nvidia Sentry, an out-of-band watchdog for in-silicon agent security.
CEO Jensen Huang said: “[AI’s] full promise can only be realised when people have confidence that AI is being built to be safe and deployed with wisdom and responsibility. This is bigger than a single product. It's the beginning of an open ecosystem to build the trust layer for safe agent systems.”
In a press release, Nvidia said the platform’s more than 100 partners would work to integrate it into their own services or expand on its abilities. Nvidia has also made OpenShell capable of running on Arm or Intel compute platforms.