Overview
- Nvidia has launched an Open Secure AI Alliance that includes the NOOA agent project to harden autonomous software and the Halos framework to reduce safety risks when AI controls physical systems.
- Cisco published a detailed Integrated AI Security and Safety Framework that catalogs attacker goals and links more than 150 techniques to standards such as NIST and OWASP to help regulated customers show compliance.
- CrowdStrike is expanding into AI-focused offerings with Falcon Cloud Security AI-SPM for continuous posture management and Frontier services for real-time model risk monitoring inside enterprise environments.
- There are no publicly reported joint guidelines tying the three firms’ programs together, so each vendor is pushing its own controls and tooling rather than a shared industry standard.
- The split approaches speed the availability of tools but could leave organizations with uneven protections and raise pressure for regulators or industry groups to force interoperable standards.