Nvidia-Led Open Secure AI Alliance Unveils Cybersecurity Framework for AI Systems

Nvidia established the Open Secure AI Alliance (OSAA), which has grown to over 120 members including Adobe, Cisco, Intel, Microsoft, and Visa, to address AI security challenges. Within a week, the group formed the Shared AI Findings Exchange (SAFE) working group and unveiled initial proposals for safely handling AI cybersecurity incidents, incident reporting, and post-incident analysis. The Linux Foundation manages the proposals. Members are contributing open-source tools, including Nvidia's Garak vulnerability scanner and Red Hat's governance solutions. Notably, Anthropic, OpenAI, and Google have not yet joined, despite OpenAI and Google signing the original open letter that prompted the initiative.