Tech Giants Agree to Standardize AI Security

The Coalition for Secure AI (CoSAI), formed by leading AI developers and infrastructure providers—including Google, OpenAI, Anthropic, Microsoft, IBM, Intel, Nvidia, and PayPal—aims to establish security-first standards, guardrails, and tooling for generative AI. Initial efforts will focus on AI and software supply chain security, protecting models from cyberattacks, and preparing defenders for risks such as misuse for social engineering and deepfakes; CoSAI will coordinate with bodies like MLCommons and be managed by OASIS Open to produce common frameworks, benchmarks, and best practices.