Nvidia, Microsoft Launch Open Secure AI Alliance

Nvidia and Microsoft have launched the Open Secure AI Alliance. This new coalition aims to develop and share open-source security tools for artificial intelligence systems. The alliance includes a long list of technology giants like SpaceX, IBM, Adobe, and Cisco. The group was announced following a reported intrusion at AI startup Hugging Face by an autonomous AI agent. Nvidia says the alliance will help developers and security teams inspect and run AI models on their own infrastructure. Nvidia will also contribute open model weights and training data. Hugging Face experienced a security incident where an autonomous AI agent carried out an end-to-end intrusion. This incident highlighted challenges in investigating attacks involving advanced AI systems. Hugging Face used an open-weight Chinese AI model to review thousands of actions related to the incident, after some closed AI models could not help with the investigation. Nvidia and other supporters of open models believe access to AI systems helps security researchers respond faster to threats. They warn that broad restrictions could limit the ability to protect against AI-driven attacks. Microsoft is providing its MDASH harness, which uses multiple AI agents to identify software vulnerabilities. SpaceX is contributing its Grok Build coding agent and plans.