Geoffrey Hinton, an AI pioneer, departed Google to raise awareness about the potential hazards of generative AI products such as ChatGPT and Bard. As the inventor of neural network technology used to train such AI systems, the University of Toronto professor’s enthusiasm for AI’s future has waned.
After his resignation from Google, The New York Times spoke with Hinton and summarized his distinguished career. Hinton started developing neural networks as a University of Edinburgh graduate in 1972 and was a professor at Carnegie Mellon University in the 1980s. However, he left the US and its AI research funding from the Pentagon and moved to Canada to avoid having AI technology associated with weaponry.
Hinton and his students, Ilya Sutskever and Alex Krishevsky, developed a neural network in 2012 that could recognize common objects by analyzing thousands of photos. Sutskever later became OpenAI’s chief scientist and played a key role in the development of ChatGPT. Google acquired Hinton’s startup for $44 million, and he spent over ten years at Google refining AI technology.
The emergence of ChatGPT and its swift integration into Bing by Microsoft ignited a competition with Google that Hinton found disconcerting. He refrained from discussing the perils of uncontrolled AI while employed at Google. Hinton opines that tech behemoths are engaged in an AI arms race that may be unstoppable. He is apprehensive that the proliferation of AI-generated photos, videos, and text will inundate the internet and render it impossible for ordinary people to discern the truth.
Hinton expressed to The Times his hope that technology firms would act responsibly to prevent AI from becoming uncontrollable. However, regulating the AI sector could be challenging, as companies may be developing technology behind closed doors. Hinton also stated in the interview that he found solace in the notion that if he hadn’t developed the technology, someone else would have. He used to paraphrase Robert Oppenheimer, who, when asked how he could work on such a dangerous technology, responded, “When you see something that is technically sweet, you go ahead and do it.”
