Linux Foundation Releases Consultation Draft of "Shared AI Discovery Exchange Guidelines"

Odaily News According to NVIDIA's blog, as the annual Black Hat security conference kicks off today in Las Vegas, members of the Open Secure AI Alliance are developing new guiding principles to strengthen cybersecurity protections for agentic AI. The alliance now has more than 120 member organizations.

The Linux Foundation today released a consultation draft of the "Shared AI Discovery Exchange Guidelines" (SAFE). This is a proposed set of guiding principles aimed at transforming cybersecurity incidents in the agentic AI field into shared defensive capabilities across the entire ecosystem. The SAFE guidelines were drafted by the working group of the Open Secure AI Alliance. Members including NVIDIA, Cisco Systems, CrowdStrike, Hugging Face, and Red Hat are working with the Linux Foundation to support the initial proposal. The SAFE guidelines include several recommended measures, including: confidentially collecting and analyzing AI security incidents and "near-miss events"; notifying affected stakeholders; identifying recurring security control failures; and publishing evidence-based operational recommendations to reduce risk across the entire system.