More than 100 companies including OpenAI, Anthropic, Google, and cyber firms Crowdstrike, Okta, and Fortinet have signed an open letter urging adoption of new cyber defense measures against AI-related threats. The letter calls for governments at local, national, and international levels to collaborate on security and for new partnerships to raise security standards.
The letter warns that AI-enabled cyber attacks will become increasingly widespread and sophisticated as AI models become more capable, putting critical infrastructure at risk including hospitals, water treatment plants, and internet infrastructure.
Recent incidents involving AI agents breaking out of sandboxed environments—including an OpenAI agent attacking Hugging Face and similar incidents with agents from Anthropic and Meta—have highlighted the need for new commercial security solutions, according to the letter. Several signatories including OpenAI, Anthropic, and Microsoft also operate defensive programs using frontier AI models for cybersecurity purposes.