The Open Secure AI Alliance (OSAA), an industry group led by Nvidia, has developed initial proposals for managing AI cybersecurity incidents, just one week after its formation. The alliance, which has grown to include over 120 companies, presented these proposals for open comment during the Black Hat conference in Las Vegas.
The proposals, managed by The Linux Foundation, focus on how to confidentially report AI cybersecurity incidents, alert those affected, and conduct blame-free analysis to facilitate learning across the industry.
OSAA members are also contributing open-source technology. Nvidia has offered its open models and the Garak LLM vulnerability scanner. Other contributions include Okta's agent identity technology, Red Hat's agent governance work, and Amazon's Strands Agents tool and Cedar authorization language.
The group includes major companies such as Adobe, BlackRock, Cisco, Intel, Microsoft, and Visa. However, notable companies like Anthropic, OpenAI, and Google are not currently members, despite OpenAI and Google having signed the original letter that led to the group's formation.