‘We have no experience with what it’s like to have things smarter than us.’ Credit: Johnny Guatto/University of Toronto
Geoffrey Hinton, one of the so-called ‘Godfathers of AI’ has won a Nobel Prize for pioneering the very same technology he fears could result in the “end of people.” Hinton and fellow AI researcher John J. Hopfield were awarded the Nobel Prize in Physics on Tuesday for their crucial early work on artificial neural networks, which have since formed the foundation for powerful AI models developed by Google, OpenAI, and others. The pair’s work, which draws on inspiration from the human brain’s architecture, paved the way for advancements in machine learning used in everything from fraud detection to driverless vehicles. In Hinton’s view, the tech he helped pioneer may also pose a profound risk to human safety.
[ Related: ‘Godfather of AI’ quits Google to talk openly about the dangers of the rapidly emerging tech ]
Early AI researchers looked to the human brain for inspiration
Hopfield and Hinton are credited with advancing the study of neural networks in the 1970s and ‘80s during a time when it was still unclear that the field would mature into the behemoth it is today. Hopfield is known for combining findings from psychics, biology, and neuroscience to create a network—the “Hopfield network”—capable of saving and recreating patterns from data. Hinton later built off of this to create his own network called the Boltzmann machine which can identify patterns in large masses of data. Together, these advancements were crucial first steps to eventually creating machines capable of classifying images. That, in turn, would be used in more modern AI models to rapidly learn from images and patterns stored on vast datasets.
The award further highlights the growing parallels and interconnectedness between psychics and computer science. Neural networks, which draw inspiration from how the human brain uses neurons to take in new information, have since gone on to form the underlying technological basis for large language models like ChatGPT, as well as image recognition models used in everything from cancer screening to facial recognition. Hinton and several of his colleagues previously received the Turing Award for their work on neural networks.
In a statement describing its decision, the Nobel committee credited Hopfield and Hinton’s work introducing new “way[s] for us to use computers” to solve challenging societal questions.
“Thanks to their work humanity now has a new item in its toolbox, which we can choose to use for good purposes,” the Nobel Committee wrote on X, the social media platform formerly called Twitter. “Machine learning based on artificial neural networks is currently revolutionising science, engineering and daily life.”
Hinton, who has previously referred to modern AI tech as an “existential threat” to humanity, appeared surprised by the Royal Swedish Academy of Science’s decision speaking with The Washington Post Tuesday morning.
“I’m in a cheap hotel in California which doesn’t have a good internet or phone connection. I was going to get an MRI scan today, but I think I’ll have to cancel that,” Hinton said. He went on to voice caution around future AI development, though his tone was markedly less pessimistic than some of his previous statements.
“It’s going to be wonderful in many respects,” Hinton told the Post, “It’ll mean huge improvements in productivity. But we also have to worry about a number of possible bad consequences … I am worried that the overall consequence of this might be systems more intelligent than us that eventually take control.”
“We have no experience with what it’s like to have things smarter than us,” he added.
Hinton left Google to speak freely about his AI fears
Google approached Hinton in 2012 after he and several colleagues achieved a breakthrough in neural network programming. Hinton joined the company as a researcher and vice president and worked to advance their AI efforts. The technology progressed much faster than Hinton was expecting. In previous reports, Hinton said he thought the types of text outputs generated by ChatGPT and other large language models in recent years would not have been possible in his lifetime. The researcher became increasingly uncomfortable with the speed of AI’s development, leading him to resign from Google last year so he could criticize the industry more freely.
After leaving Google, Hinton joined a chorus of alarmed researchers and technologists who believe the tech industry’s rapid rush to create more powerful AI models could have harmful societal side effects. Although he’s since tempered his language in recent statements, Hinton previously expressed concerns that an unchecked AI model could, somehow, “take over” humanity. Those fears caused the researcher, known by many in the industry as a “Godfather” of AI, to reportedly regret much of his life’s work.
“The idea that this stuff could actually get smarter than people—a few people believed that,” Hinton said during a 2023 interview with The New York Times. “But most people thought it was way off. And I thought it was way off. I thought it was 30 to 50 years or even longer away. Obviously, I no longer think that.”
Hinton’s new Nobel Prize may further complicate the researcher’s already strained relationship with his legacy.
Source : Popular Science