IT Home On October 17, according to Reuters, Demis Hassabis, co-founder and chairman of Google DeepMind, said on Wednesday that recent cyberattack incidents initiated by AI agents are “early warnings”. If systems do not further improve their security, the world may soon face serious cybersecurity issues.
Hassabis, who is also the chief scientist at Google’s parent company Alphabet, said at an artificial intelligence summit in Dublin: “I believe these attack incidents have sounded a warning bell for us in the early stages.“The current system is not powerful enough, so the damage caused is still limited.”
But looking ahead to the next one to two years, if all industries, companies, and governments do not invest more in protecting these systems, it is very likely to turn into a catastrophic major event in the future.
IT Home noticed that Microsoft President Brad Smith pointed out at the conference that the uncontrolled behavior of AI has made 2026 a “warning year”. There is an urgent need for new AI security laws and regulations to ensure that safety measures can keep up with innovation.
Smith emphasized: “No company or technology can override the law. But we need laws and regulation. This regulation should be reasonable, balanced, and take innovation into account. To be honest, the resistance among some in the industry to regulation that they fear is an obstacle we must overcome.”
Hassabis added that although establishing an international standard body seems like a logical intervention, the mechanism must be balanced and responsive to avoid stifling technological innovation.
Therefore, the key lies in dynamically building a certification mechanism that can accurately match the actual level of risk and can be quickly adjusted as technology evolves.