Five Eyes Warns: AI to Overwhelm Cybersecurity in Months

Five of the world's most powerful intelligence agencies, known as the Five Eyes alliance, warn that frontier AI models will overwhelm existing cybersecurity defenses in months, not years. This timeline reflects capabilities that already exist. What's interesting is that Anthropic’s Mythos-class AI models have shown they can break down cyber defenses and identify software flaws. In one demonstration, the Mythos model penetrated almost all classified systems at a sophisticated intelligence agency within hours. The Five Eyes advisory highlights that these AI models lower the barrier for malicious actors to find vulnerabilities and execute sophisticated attacks at a speed and scale existing defenses were not designed to handle. The threat is not new attack methods, but that more capable AI makes existing weaknesses much easier and faster to exploit. The bottom line is that defenses adequate today may not hold in six months because attacker tools will have quietly improved beyond them. This means organizations need to confront this central tension immediately.