NVIDIA has joined the Open Secure AI Alliance, a collaborative effort to improve cybersecurity through open models and tools. The alliance aims to ensure that defenders have access to open, frontier technologies to safeguard software and agents in the age of AI. The initiative builds on the Linux Foundation’s Akrites initiative and OpenSSF community work, focusing on remediation and disclosure of vulnerabilities using open technologies. According to NVIDIA, the alliance will enable a multi-vendor ecosystem for critical industries to build security systems and avoid single points of failure. The recent Hugging Face security incident demonstrated the importance of open systems for self-defense, as the company used an open-weight model to analyze and contain an intrusion. Source: nvidia
NVIDIA is contributing open models, model weights, data, and new agent harness research to the alliance to accelerate the development of cybersecurity tools and techniques. The new open source NVIDIA Labs Object-Oriented Agent (NOOA) project is now available on GitHub, making advanced AI safety capabilities more accessible for agent harnesses. The NOOA research framework enables harnesses to better integrate with models, making agent behavior easier to test, trace, audit, and govern. Across the alliance, contributors are building an open defense stack for agents, covering identity, isolation, safe model formats, multi-model scanning, and secure coding workflows. Other members include HPE, Hugging Face, IBM, Microsoft, and SpaceXAI, each contributing tools and technologies to enhance open security. Source: nvidia
The alliance emphasizes that open models and tools are essential for cybersecurity, as they democratize defensive capabilities, increase transparency for defenders, and enable cyber defense while protecting data. The alliance argues that risks associated with open systems are not unique to them and must be managed wherever advanced AI is deployed. The Open Secure AI Alliance is a movement to develop and share open technologies, techniques, and tools to safeguard software and agents in the age of AI. Source: nvidia
