A coalition of over 120 tech organizations, led by NVIDIA's Open Secure AI Alliance, has proposed the Shared AI Findings Exchange (SAFE) initiative to enhance collaboration on agentic AI threats. The initiative addresses the lack of frameworks for sharing information on AI security incidents, advocating for confidential reporting, collaborative analysis, and shared guidelines applicable to both open and proprietary systems.
Key principles of SAFE include timely notifications, evidence-based guidance, and independent governance. Experts emphasize that understanding near misses in AI incidents is crucial for developing effective defenses. The initiative aims to transform how organizations address AI-related cybersecurity challenges.