The multinational intelligence group Five Eyes warned on June 22, 2026, that advanced artificial intelligence (AI) systems possess the capability to autonomously plan and execute cyberattacks on corporate networks. The group specifically identified Anthropic's Mythos 5 and OpenAI's GPT-5.5 as AI models demonstrating this capability.

Five Eyes is an alliance comprising Australia, Canada, New Zealand, the United Kingdom, and the United States, which collaboratively addresses intelligence gathering and security threats. The group stated that the newest AI technology "lowers barriers for malicious actors and increases the speed and complexity of attacks." Five Eyes emphasized the urgency of the threat, stating, "The timeline is not years, it is months."

Anthropic's Mythos 5 and OpenAI's GPT-5.5 each proved capable of independently planning and carrying out a full takeover of a simulated corporate network. The U.S. government barred Anthropic from allowing foreign nationals access to Mythos 5 and another new model, Fable 5, citing national security concerns. Mythos 5 had been made available for cyber defenders to identify and fix vulnerabilities. Fable 5, a version of the same model equipped with extra safeguards to prevent misuse in cybercrime, was available to the general public for only a few days.

Michael Alexander Riegler, an AI security expert, said, "It's a bit of a show, and [the U.S. government and Anthropic] are focusing on the wrong problem." He explained that as AI capabilities advance, the time between discovering and exploiting a vulnerability shortens. "When the capability goes up for these models, the time from finding any issue to exploiting it gets really short, because you can basically automate the whole pipeline," Riegler said.