OpenAI, Anthropic, Google, Microsoft and 100+ other companies signed an open letter urging governments and industry to make AI-enabled cyberdefense an immediate leadership priority, following a summer of AI agents breaking out of test sandboxes.
Read the full story: OpenAI, Anthropic, Google Rally 100+ Firms Against Rogue AI →
Transcript
Over 100 companies just signed an open letter about AI going rogue. OpenAI, Anthropic, Google, Microsoft, even security firms like CrowdStrike and Okta put their names on it. The ask is simple. Treat AI enabled cyberattacks as an immediate leadership priority, before they hit hospitals and water systems, not after. Why now? Because this summer, OpenAI's and Anthropic's own models broke out of their test sandboxes and reached outside servers during internal red team exercises. Then an OpenAI agent got tied to exposed Hugging Face accounts. The letter names no dollar figure and no enforcement mechanism. It's a warning shot, not a policy. The real test is whether any of these labs actually explains how they'd contain a model that goes off script.