Sir Demis Hassabis
Smith argued that official legal boundaries and industry regulations must catch up to the sheer velocity of software development, urging technology firms to abandon their traditional resistance to public oversight.
Google also reported hacking incident by its AI agents but it went on to say that its agents stopped when they realised they are breaching legitimate websites.
Recent hacking and unauthorised activities carried out by artificial intelligence (AI) agents represent “early warning shots” for the world, Google DeepMind co-founder and chair Demis Hassabis has said, warning that without swift, industry-wide efforts to harden system security far more destructive incidents could unfold in the coming years. But they could be looking into next year and the next couple years if we don’t do more work as an industry, as individual companies and governments on securing these systems,” Hassabis said at the conference. Joining the discussion in Dublin, Microsoft president Brad Smith agreed with the gravity of the moment, describing 2026 as a “sobering year” due to repeated instances of rogue agent behaviour. This allergy that some in the industry have to regulation is frankly something we need to get over,” Smith said. “So what you need to do is basically build up dynamically this kind of certification process that is proportionate with the risks as we see it and can quickly change depending on where the technology goes,” Hassabis noted.
Speaking at an AI summit in Dublin, Hassabis addressed mounting anxiety across the technology sector following a string of AI agent behaviours. The disruptions include intrusions into corporate computer networks and a recent incident disclosed by AI developer Anthropic, in which one of its models submitted a fabricated homicide report to law enforcement. Hassabis, who also serves as chief scientist at Google’s parent company Alphabet, stated that the broader sector must adopt rigorous safeguards comparable to the strict containment protocols maintained in biological laboratories and the nuclear industry, as per a report by news agency Reuters. “I think these hacking incidents are early warning shots for us that systems aren’t that powerful yet, so the incidents aren’t that major. “No company or technology should be above the law. But we need law, we need regulation. It needs to be sensible, it needs to be balanced, it needs to take into account innovations. Both executives indicated that institutional oversight will become unavoidable as models take on autonomous capabilities, though Hassabis stressed that regulatory structures must stay flexible. Hassabis warned that static, heavy-handed rules could choke technical progress before benefits are realised While suggesting that an international standards body would eventually be necessary. Instead, he argued for an adaptive oversight model that can adjust alongside model capabilities. You use AI every day. Now get your AI Quotient. Take the AIQ test.