Professor Hinton, Interview with Japan's Yomiuri
ChatGPT Can Accumulate Knowledge Thousands of Times That of the Human Brain
AI May Surpass Human Abilities as Early as the 2020s
"Generative artificial intelligence (AI) may surpass human intelligence and have the potential to dominate human society."
Professor Geoffrey Hinton of the University of Toronto, known as the 'Father of AI,' stated this in an interview with Japan's Yomiuri Shimbun on the 4th, saying, "I do not know how humanity can respond to the threat of having human society taken over by digital beings with intelligence higher than their own."
He analyzed that 'ChatGPT,' released by OpenAI, which has recently experienced CEO ousting and reinstatement amid conflicts over AI risks, can already accumulate knowledge thousands of times that of the human brain. He predicted that AI could surpass human abilities in several areas as early as the 2020s.
Professor Hinton cited manipulation of public opinion by authoritarian governments and election interference using fake news as risks posed by AI. He said, "In the United States, one of the major political parties is allowing the spread of false information to gain an advantage in election campaigns," and added, "There is a need for legislation regulating the creation and possession of fake videos."
He also warned that AI-equipped weapon systems could pose concrete dangers. He expressed concern that if systems that automatically set attack targets using AI become practical, wars could become uncontrollable. He had previously mentioned, "I fear the day when AI turns into killer robots." Furthermore, he predicted that AI replacing human jobs would lead to an expansion of the wealth gap in the future.
Above all, Professor Hinton urged government agencies to take direct action to minimize AI risks. He emphasized, "It is very difficult to regulate because large corporations and countries are competing in development," and "Current measures are insufficient. Humanity needs a change in direction." He appealed for governments worldwide to cooperate and put brakes on the global AI development race.
Professor Hinton, who has recently issued repeated warnings about AI risks, is called the Father of AI for advancing deep learning technology, the core of AI. He played a key role in AI technology development while working at Google for the past decade. However, after Google launched the AI chatbot 'Bard' in March this year to compete with ChatGPT, he suddenly resigned from Google in April, just a month later. He explained that the reason for parting ways with Google was "to freely criticize the risks of AI technology." Subsequently, Professor Hinton likened AI to a "nuclear weapon" and expressed regret over dedicating his life to AI development.
On the 17th of last month (local time), Professor Hinton was also the academic advisor of Alia Sutskover, the chief scientist and a board member who led the dismissal of OpenAI CEO Sam Altman. Sutskover pursued her PhD in computer science under Professor Hinton at the University of Toronto. She worked at DNN Research, an AI company founded by Hinton and acquired by Google in 2013, and was greatly influenced by him. As concerns over Altman's aggressive AI development and commercialization were reportedly significant among OpenAI board members including Sutskover, attention has also been drawn to the connection between Professor Hinton and Sutskover, who have both raised voices of caution regarding AI risks.
© The Asia Business Daily(www.asiae.co.kr). All rights reserved.
