100+ firms sign rogue-AI cyber pledge

More than 100 companies, including OpenAI, Anthropic, Google, Microsoft, and cybersecurity firms Crowdstrike, Okta, and Fortinet, signed an open letter August 27 urging private and public-sector cooperation to defend against AI-enabled cyberattacks. The letter warns:

AI-enabled cyber attacks will become far more widespread and sophisticated as models around the world become increasingly capable.

and flags critical infrastructure like hospitals and water treatment plants as particularly exposed. It calls for new partnerships between industry and governments at the local, national, and international level to raise security standards, without naming specific mandates. The letter references each signer’s own defensive tooling by name, including OpenAI’s Daybreak, Anthropic’s Mythos, and Microsoft’s Perception, as examples of the kind of monitoring systems it wants to see more widely adopted and shared across the industry.

What it means for you

This follows a run of incidents this year, including the agentic breach traced back through Hugging Face’s platform, that made rogue-model cybersecurity a boardroom topic rather than a research concern. An open letter with no binding mandate doesn’t change your risk exposure today, but naming each vendor’s internal defense system publicly, Daybreak, Mythos, Perception, is itself useful: it tells you these systems exist and gives you a term to ask your vendor about when you’re evaluating how seriously they take an incident like the one Hugging Face lived through.