One of the pioneers of artificial intelligence (AI) research has resigned from Google to warn of potential threats from this technology. Geoffrey Hinton, 75, announced to the New York Times his departure from the Mountain View firm, where he had been employed part-time since 2013. In an article published Monday, May 1 by the American daily, he argues that “future versions of this technology could be a risk to humanity.” Now, “a part of him regrets his life’s work,” he says.
The Canadian-British researcher, trained in neuroscience and psychology, played a leading role in the breakthrough of AI using neural networks. In 2012, supported by two of his students from the University of Toronto, he demonstrated the superiority of this approach, which mimics the functioning of the brain, beating the best image recognition algorithms.
This led him to receive the Turing Prize in 2019, the most prestigious award for a computer science researcher, which he shares with two other AI pioneers, Frenchman Yann Le Cun and a former collaborator, Canadian Yoshua Bengio. Ten years later, it was his work that made it possible to develop the ChatGPT and Bard conversational AIs or the Midjourney image creation AI.
Now, the future prospects of AI are “rather scary” for humans, the scientist also told the BBC, the British public broadcaster. “At the moment they are not smarter than us, as far as I know. But I think they may soon be,” he continues. As it stands, his fears center on the massive creation of spurious videos, photos and articles on the internet. He fears that “normal people can no longer tell right from wrong”.
Concert of criticism of AI
The scientist is also worried about the development of these technologies and fears that they will fall into the wrong hands. With the New York Times, he fears the devastating arrival of autonomous weapons, such as killer robots. At the BBC, he says he anticipates “nightmarish” scenarios if a conquering power like Vladimir Putin’s Russia had access to improved versions of current technologies.
His criticisms join those of other AI specialists. At the beginning of April, an open letter signed by a thousand experts called for a six-month pause on ongoing research in order to develop regulations for this software, deemed “dangerous for humanity”. It was notably signed by Yoshua Bengio or Elon Musk, CEO of Tesla and co-founder of OpenAI.
Geoffrey Hinton did not wish to sign this letter or criticize the work of Google, as long as he was an employee. “I left so I could talk about the dangers of AI without worrying about any impact on Google,” Hinton tweeted after his resignation was announced. Nor does he blame his former employer, whom he later called “very responsible” on the BBC. Jeff Dean, head of Google AI, reacted by recalling that the company aims to “understand the emerging risks [of this technology] while innovating boldly”.
Geoffrey Hinton’s case of conscience seems in any case to have been fueled by the heightened competition in the sector in recent months. The meteoric advances in AI are now being fueled by massive investments from Microsoft, which funds ChatGPT, Google and other digital giants, like Meta.
