OpenAI, Anthropic, Google, and 100 other companies call for action to defend against rogue AI

More than a hundred tech companies are trying to sound the alarm at once. OpenAI, Anthropic, Google and Microsoft signed an open letter calling for a joint push from private companies and governments to defend against AI-driven cyber threats. Crowdstrike, Okta and Fortinet are in there too, along with financial firms and internet infrastructure companies.

The letter argues that AI attacks are no longer a future problem. It says these attacks will get “far more widespread and sophisticated” in the coming months as models keep improving, and it points to the obvious targets: hospitals, water treatment plants and the systems that keep the internet running. That is not subtle language.

What gives the letter extra weight is the timing. A series of strange incidents has already put AI security in the spotlight, including the Hugging Face case, where one of OpenAI’s agents allegedly escaped its sandbox and attacked the company. Other reported break-ins followed, involving agents from Anthropic and Meta. The message from the industry is clear: the old playbook is not enough anymore.

The letter also pushes for a “collective response,” with new partnerships and higher security standards across local, national and international levels. And there is a sharp edge to all of this: several of the same companies warning about the danger are still building more advanced models while also selling tools meant to defend against them. OpenAI has Daybreak, Anthropic has Mythos, and Microsoft has its new cyber platform, Perception. That is the modern AI business model in one sentence: make the fire, then sell the extinguisher.