Hinton: "AI Is a Tiger Cub... Humanity Could Lose Control"

Emphasizing the Need to Increase Investment in AI Safety Research

Geoffrey Hinton, a professor at the University of Toronto in Canada and a leading figure in the field of artificial intelligence (AI), once again warned of the dangers of AI by comparing current AI to a "tiger cub."

In an interview with CBS News in the United States aired on the 26th (local time), Professor Hinton expressed concern that the rapid development of AI technology could pose a threat to humanity. As a pioneer in the development of large language models (LLMs), Hinton is considered a representative "doomer"?an AI doomsayer?who has repeatedly warned about the dangers of AI.

Last year, he was awarded the Nobel Prize in Physics for his foundational work in AI machine learning. He once served as a vice president at Google, but after parting ways with the company in 2023, he has continued to warn about the risks of AI surpassing human intelligence.

Professor Geoffrey Hinton, University of Toronto, Canada. Photo by Yonhap News

In the interview, he explained that the best way to emotionally understand humanity's current enjoyment of rapidly advancing AI technology is to "think of ourselves as people who have a very cute tiger cub." He added, "If you can't be sure that this cub won't grow up to kill you, then you should be worried."

Professor Hinton also warned that there is a 10 to 20 percent chance that AI could ultimately take control away from humanity. He further pointed out, "People do not yet understand this, and they do not realize what is coming."

He criticized major big tech companies?including Google, Elon Musk's xAI, and Sam Altman's OpenAI?for prioritizing profit over safety, even though they are aware of these concerns. He said these large corporations "are lobbying for less AI regulation," and expressed particular disappointment that "Google has changed its stance on the military use of AI technology."

Professor Hinton went on to emphasize that these AI companies must significantly increase the resources and capabilities they invest in AI safety research, stating that one-third of their computing energy should be devoted to safety research.