Coalition for Secure AI Releases Two Actionable Frameworks for AI Model Signing and Incident Response

Press Release:
Boston, MA – 18 November 2025 – OASIS Open, the international open source and standards consortium, announced the release of two critical publications advancing AI security practices from the Coalition for Secure AI (CoSAI), an OASIS Open Project. These new resources provide practical frameworks to help organizations strengthen the security and trustworthiness of their AI systems. CoSAI’s Software Supply Chain Security for AI Systems Workstream released “Signing ML Artifacts: Building towards tamper-proof ML metadata records” and the Preparing Defenders for a Changing Cybersecurity Landscape Workstream published “AI Incident Response Framework V1.0.” Together, these frameworks address key aspects of the full lifecycle of AI assurance, from preventing tampering before deployment to responding effectively when systems are attacked.

Read the full press release published on the OASIS website here.