Nvidia-Led Industry Group Launches AI Security Working Group Amid US Fears Over Chinese Models

The Open Secure AI Alliance (OSAA), an industry coalition led by Nvidia, has formed a new working group to develop AI security guidelines. The Shared AI Findings Exchange (SAFE) working group presented initial proposals for open comment this week at the Black Hat cybersecurity conference in Las Vegas. These proposals focus on confidential incident reporting, alerting affected parties, and blame-free analysis following AI security incidents.

Alongside these proposals, OSAA members are cataloging open source security tools, including Nvidia's LLM vulnerability scanner Garak, agent identity work from Okta, agent governance efforts from Red Hat, and open source contributions from Amazon. The group's membership includes prominent companies like Adobe, BlackRock, Cisco, Intel, Microsoft, Visa, and Hugging Face.

Notably absent from the coalition are Anthropic, OpenAI, and Google, despite their earlier support for open source AI development through an open letter championed by Nvidia. The formation of OSAA follows recent reports that U.S. officials considered restricting Chinese open weight AI models, sparking industry pushback and calls for greater openness as a security strategy.