DeepMind co-founder: AI cyberattacks are a warning of worse to come

Google DeepMind co-founder Demis Hassabis said at an AI summit in Dublin, Ireland that recent cyberattacks launched by AI agents are an early warning. Without better system security, the world could soon face serious cyber incidents. He noted current systems are not strong enough yet, so the damage stays limited. But in the next one to two years, if companies and governments do not invest more in protection, things could turn catastrophic. Microsoft President Brad Smith said AI going out of control makes 2026 a "wake-up year", and called for new AI safety laws that are fair, balanced and open to innovation, since no company or technology should be above the law. Hassabis added that any global standards body must stay flexible and fast-moving. The key is a dynamic certification system that matches real risks and adapts quickly as technology evolves.