Google's chief scientist warns: If enterprises, governments, and others do not strengthen the security protection of AI systems, it could escalate into a major security incident.

PANews, October 15 news: At an artificial intelligence summit, Demis Hassabis, chairman of DeepMind and chief scientist at Google, warned that if enterprises, governments, and other parties do not strengthen the security protection of AI systems, next year or in the coming years it could escalate into a major security incident. Hassabis believes that over the past year the AI industry has entered the era of agents. AI being able to act autonomously means it has the opportunity to complete more real-world tasks, and its practical value has accordingly increased. But on the other hand, the potential risks of the systems are rising. This year's AI attack incidents are a signal, but because AI capabilities are not yet considered powerful, the accidents have not yet caused major consequences. At the summit, Hassabis called on the AI industry to draw on the mature experience of high-risk fields such as biology and nuclear technology to establish stricter security assurance mechanisms, and said there is "still a great deal of work to be done" in this regard. He emphasized the importance of international cooperation and proposed that in the future it will be necessary to establish an international standards body to formulate safety specifications for AI systems.