'AI Daebu' Hinton: "Safe AI Development Becomes Difficult if Trump Takes Office"

Warning on AI Safety in Yomiuri Shimbun Interview
Nobel Prize in Physics Awarded This Year for Contributions Based on Machine Learning

This year’s Nobel Prize in Physics laureate, 'the godfather of Artificial Intelligence (AI)' Geoffrey Hinton, Professor Emeritus at the University of Toronto, warned that under the Donald Trump administration set to launch in January next year, "safe AI development will become more difficult."

In an interview with the Yomiuri Shimbun published on the 3rd, Professor Hinton expressed concern over Trump’s declaration to abolish the executive order on AI safety introduced by the Joe Biden administration, saying, "(The Trump administration) will allow AI companies to freely create their own norms."

Reuters Yonhap News

Professor Hinton stated that the current capacity devoted by AI development companies to safety research is at the 1% level and emphasized, "Investment in safety research needs to be increased thirtyfold." He further pointed out, "It is necessary to revise laws so that AI development companies allocate one-third of their computing power to safety research."

Professor Hinton received this year’s Nobel Prize in Physics along with John Hopfield, Professor at Princeton University, for their contributions to establishing the foundation of machine learning through research on artificial neural networks that ushered in the current AI era. After conducting research at Google for many years, he resigned last year and has continuously voiced concerns about AI risks.

Regarding the entry of advanced US IT companies into AI development, Professor Hinton criticized, "When profit-seeking companies compete, safety takes a backseat," calling it "essentially an arms race." Evaluating China’s AI development, he said, "China is investing more funds than the US and is likely to surpass the US within 10 years." He also emphasized that Japan has "very excellent researchers" and highlighted that fields such as robots for elderly care are areas where Japan can hold an advantage.

Professor Hinton also spoke about the threat AI poses to jobs. He predicted that while AI advancements could reduce misdiagnoses in the medical field and decrease accident risks in autonomous vehicles, the number of jobs lost to AI will exceed the number of new jobs created by it. Furthermore, Professor Hinton estimated that the probability of AI surpassing humanity within 20 years is over 50%, and it could emerge as soon as five years from now. He advised, "We need to seriously consider the possibility that AI could dominate humanity."

This content was produced with the assistance of AI translation services.