NVIDIA has brought together the world's leading technology companies to create the Open Secure AI Alliance, which aims to improve cybersecurity. The 27 founding members include Microsoft, SpaceX, Dell, The Linux Foundation, and NVIDIA itself.
NVIDIA has brought together the world's leading technology companies to create the Open Secure AI Alliance, which aims to improve cybersecurity. The 27 founding members include Microsoft, SpaceX, Dell, The Linux Foundation, and NVIDIA itself.
“The Open Secure AI Alliance, building on the leadership of the Linux Foundation’s Akrites initiative and the work of the OpenSSF community, will work to address and disclose vulnerabilities using open technologies,” NVIDIA said, according to Engadget.
The alliance members argue that to effectively defend against attacks in the era of AI models like Anthropic’s Mythos 5, security researchers need access to both open and proprietary advanced models. NVIDIA cited OpenAI’s recent audacious attack on Hugging Face as a prime example. Hugging Face initially tried to use proprietary, commercial advanced models to detect and repel attacks, but its queries were blocked by defenses. The company then switched to open-source models and successfully stopped the attack.
“When closed AI tools, unable to distinguish attackers from defenders, blocked critical digital analysis, Hugging Face deployed a model with open weights GLM 5.2 on its own infrastructure to analyze over 17,000 actions and localize the intrusion,” NVIDIA noted.
NVIDIA and several of its partners are joining the initiative, providing various components to the alliance. NVIDIA said it will provide open models, model weights, data, and new agent development environments to accelerate the creation of new cybersecurity tools and techniques. HPE will contribute standards and methods for cryptographically validating AI agents and services, while Hugging Face will provide a secure format for storing AI model weights called Safetensors. Microsoft, SpaceXAI, IBM, and Red Hat will also provide their open-source AI technologies.
The group also calls on governments to collaborate with companies and invest in a shared open AI infrastructure. The alliance added that regulators should view open models as “assets to protect, not threats.” They stressed that blanket restrictions on open advanced AI systems “weaken defenses and risk concentrating power” in the hands of a few closed vendors.
