Google DeepMind’s head calls for establishing a US-led global AI regulatory body

DeepChao TechFlow message: On July 14, according to AXIOS, Demis Hassabis, co-founder and CEO of Google DeepMind, called on the United States to establish a new AI regulatory body with the authority to review the world’s most advanced AI models and to coordinate an industry-wide slowdown in development if risks intensify. Hassabis said it is time to take a more “systematic” approach to AI regulation—industry-funded, staffed by top-tier technical experts worldwide, and accountable to the U.S. government. He said the current cybersecurity risks posed by AI are only “early warning signals.”

He believes that within the next 18 months, these capabilities—and even more severe bio and nuclear security risks—could emerge in open-source models that no government can control. He also emphasized that the risks will not only come from open-source models, but also from the stronger proprietary models that major AI labs may roll out in the future. It is reported that Hassabis has been quietly pushing this plan for months; before any public announcement, he has communicated with the Trump administration, leaders of other AI labs, and European officials, and said that the feedback he has received has been “very positive.” (Jin10)