Despite his graying hair, Geoffrey Hinton still looks top of the class. Navy blue round-neck sweater, pink shirt that barely sticks out and that still surprised look, as if he had just made a new discovery. He is not a regular in the media. Even less to express his fears related to new technologies. But that was before he left Google, his parent company for more than a decade. “I left so I could talk about the dangers of AI without worrying about the possible impact on Google,” he said in a tweet, after news of his departure from the The New York Times.
If he refuses to criticize the Tech giant, he is now playing the whistleblower. And this man has a heavy heart, according to what we can read in an interview published this Monday, May 1, by the American daily. The British researcher, who emigrated to Canada, talks about the potential risks of artificial intelligence (AI) and his regrets about his role in its design: “I console myself with the normal excuse: if I hadn’t done it, someone else would have done it.”
Geoffrey Hinton is nicknamed “the godfather” of artificial intelligence. And for good reason: for half a century, the researcher has developed the technology at the heart of chatbots like ChatGPT. “Few people can speak about artificial intelligence with such authority”, points out the Echo of Belgium. He was one of the pioneers of artificial intelligence, a set of theories and techniques implemented to produce machines capable of simulating human intelligence. For some, this shift may seem surprising, but when we sift through the career of this computer science professor, he has always been marked by his convictions. And no question that this will change at the age of 75.
Back in 1972. At the time, Geoffrey Hinton was a graduate student at the University of Edinburgh. Already, he is developing an idea called “neural network”. What is this ? It is a mathematical system that builds skills by analyzing data. Few researchers see the benefit of this advance. Geoffrey Hinton, he decides to make his life. In the 1980s, he left Carnegie Mellon University, located in Pittsburgh, for Canada. The reason: he is reluctant to accept Pentagon funding. “At the time for most AI, research in the United States was funded by the Department of Defense,” recalls the New York Times. However, Dr. Hinton is opposed to the use of artificial intelligence on the battlefield – what he calls “robot soldiers”.
Consecration: the Turing Prize
The year 2012 marks an important turning point in the researcher’s life. Geoffrey Hinton and two of his graduate students from the University of Toronto created technology that became the intellectual foundation of AI. She is able to analyze thousands of photos and learn to identify common objects: cars, animals, plants. And their system has led to the creation of increasingly powerful technologies, including new chatbots like ChatGPT and Google Bard. Also in 2012, he created an online course on the platform Coursera in 2012 on artificial neural networks. A year later, he joined the multinational with multicolored letters: Google.
In the scientific community, the rewards are not long in coming. In 2018, Dr. Hinton and two other collaborators – including Yann Le Cun, now head of AI at Facebook) – received the Turing Prize, also called “the Nobel Prize in computing”, for their work on the neural networks. Around the same time, Google, OpenAI, and other companies began creating neural networks that learned from massive amounts of digital text. Industry leaders believe the new AI could be as big as the introduction of the web browser in the early 1990s and could lead to breakthroughs in areas ranging from drug research to education. Prospects that have long overshadowed the concerns that remain. “It’s hard to see how you can stop bad actors from using it for bad things,” Dr. Hinton said.
“Dr. Hinton thought it was a powerful way for machines to understand and generate language, but it was inferior to how humans manipulated language. Then, last year, when Google and OpenAI were building systems that used much larger amounts of data, his perspective changed,” he said. New York Times. Also interviewed by the BBC, Geoffrey Hinton said about recent developments in AI: “The problem is that it works so much better than we imagined. So what can be done to reduce the risk that in the long term , things smarter than us are taking over.” As companies perfect their AI systems, he believes, they become more and more dangerous. “Look how it was five years ago and how it is now,” he said of AI Technology. His concern is that the internet will be flooded with fake photos, videos and texts, and that the average person “will no longer be able to tell what is real”.
A possible threat to humanity
Last April, Chris Pissarides – Nobel Prize in Economics in 2010 – believed that models such as ChatGPT could help implement the four-day week. However, Geoffrey Hinton fears that AI technologies will eventually disrupt the job market. Today, chatbots like ChatGPT tend to supplement the work of humans, but they could replace personal assistants, translators and others who take on tasks by rote. “It takes the heavy work away,” he said. “It could get more out of it than that.” More generally, Geoffrey Hinton sees these new technologies as a possible threat to humanity. From vast amounts of data, they can learn to behave unexpectedly, which can become a problem, he says.
This Monday, Geoffrey Hinton joined the camp of skeptics. This statement echoes the growing number of criticisms made by Tech figures (including Elon Musk and Steve Wozniak) in recent days. Dr. Hinton had not participated in these speeches, waiting to leave his job at Google. It’s done.