The Open Secure AI Alliance, led by Nvidia, unveils its first proposals to Black Hat

A week after its formation, the industry alliance of more than 120 companies presented a mechanism for confidential sharing of AI-related security incidents.

Formed at the beginning of August around Nvidia, the Open Secure AI Alliance (OSAA) claims more than 120 member companies, including Adobe, BlackRock, Cisco, Intel, Microsoft, Visa, Red Hat, Okta, Amazon and Hugging Face. It follows an open letter signed at the end of July by more than 200 companies. Anthropic, OpenAI and Google, despite being signatories of this initial text, are not among the active members.

On August 4, at the Black Hat conference in Las Vegas, the alliance presented the Shared AI Findings Exchange (SAFE), a working group responsible for defining rules for confidential reporting of AI-related cybersecurity incidents: alerting the parties concerned, analysis without finding fault, circulation of lessons.

Several members are also contributing open source bricks to the common pot — a vulnerability scanner on the Nvidia side, agent and authorization policy components on the Amazon side, and work on agent identity on the Okta side.