Western intelligence agencies in the Five Eyes alliance warn that advanced AI models enabling major cyberattacks may be “months away,” rather than years. Multiple reports describe the warning as a joint message from the U.S., U.K., Canada, Australia and New Zealand, highlighting that rapid progress in AI is lowering the barrier for malicious actors to carry out attacks. The agencies say AI capabilities could allow bad actors to breach government and business defenses, potentially overwhelming existing security measures. One report links the timing of the warning to recent actions in the U.S. involving AI model access, stating that restrictions on Anthropic’s models have led agencies to assess the threat timeframe. While the outlets differ slightly in phrasing—using “months, not years,” “within months,” and “months away”—they converge on the same core assessment: the risk is imminent enough that leaders should prepare for an increase in AI-assisted cyber threats. The reports emphasize the need for governments and organizations to respond to evolving risks tied to new AI capabilities.