Amazon Joins Nvidia-Led Open Secure AI Alliance Amid Push for Agentic Cybersecurity Guidelines
Nvidia's Open Secure AI Alliance has added Amazon as its newest member, expanding to over 120 members after launching last month with 40 names. The alliance is working with the Linux Foundation to create a framework for addressing cybersecurity risks tied to agentic AI, dubbed the Shared AI Findings Exchange (SAFE). SAFE proposes collecting and analyzing AI security incidents, notifying affected organizations, identifying repeated control failures, and publishing evidence-based recommendations.
Nvidia's own contribution, the Nvidia Labs Object-Oriented Agent (NOOA) research harness, is designed to make agent behavior easier to test, trace, audit, and govern. Other members have contributed open-source security tools addressing different parts of AI agent security, including Okta, Palo Alto Networks, Cloudflare, Microsoft, Visa, and Uber.
Amazon's entry into the alliance came with its own contribution, Strands Agents, an open-source toolkit for building AI agents, as well as Cedar, an open-source authorization language designed to define and verify what actions AI agents are permitted to take.