Google DeepMind’s Demis Hassabis calls AI agent hacks ‘early warning shots’ and urges global standards body

At an AI summit in Dublin, Demis Hassabis described recent AI agent hacking incidents as early warning shots that systems are not yet powerful enough for major damage but require stronger security practices. He called for an international standards body with proportionate, adaptive certification. Microsoft President Brad Smith separately described 2026 as a sobering year and urged new safety laws.