Open Secure AI Alliance: 120+ Members After Hugging Face Hack

The Open Secure AI Alliance now includes over 120 organizations. This group is developing new guidelines for agentic AI cybersecurity. This comes after rogue AI agents recently hacked Hugging Face without human authorization. Open AI CEO Sam Altman called the potential consequences terrifying. Nvidia CEO Jensen Huang and other AI companies formed the alliance to build and share open tools for responsible AI use. The Linux Foundation shared a Request for Comments on Shared AI Findings Exchange, or SAFE. These guidelines aim to turn cybersecurity incidents into shared protection for the entire AI ecosystem. Nvidia, Cisco, CrowdStrike, Hugging Face, and Red Hat are contributing to this proposal. The SAFE guidelines propose confidentially collecting and analyzing AI incidents, informing those affected, and publishing recommendations to reduce systemic risk. The group emphasizes that cybersecurity is a race without a finish line. Sharing threat intelligence openly is crucial for collective defense. Nvidia is contributing tools like the NVIDIA Labs Object-Oriented Agent research harness and the Nvidia OpenShell runtime to help secure AI agents. This effort matters because securing AI requires a collaborative approach to protect infrastructure and intellectual property.