British-born, Toronto-based computer scientist and AI pioneer Geoffrey E. Hinton is known for his contributions to artificial neural network theory. In 2012, he and 2 students at the University of Toronto laid the foundation that underlies generative AI systems today.
Most recently, the 75-year-old had worked for Google after Google bought his AI startup DNNresearch for $44 million, which led to technologies such as ChatGPT and Bard. But now he has quit the job and resigned from his position. Now he joins the list of computer scientists warning about potential dangers of AI systems like ChatGPT or Bard, as reported by the New York Times (NYT).
Everything he says has already been heard from others. Such as, "It's hard to imagine how you're going to stop criminals from abusing AI to do bad things," as he tells the NYT while speaking in his hometown of Toronto. Or else, "Chat systems like ChatGPT will flood the Internet with misinformation, and people won't be able to tell what's true or what's false." But with him, considered the godfather of AI, the words definitely take on a different weight.
Hinton also points to other dangers that he thought were in the distant future. He explains that 5 years ago he had a very different view of the dangers of AI. Perhaps in 30-50 years, he thought, AI could be more intelligent than humanity and then also become dangerous. But now he has seen how AI has developed in just 5 years. It was scary, he said, to imagine where it might go in the next 5 years. "Maybe what's going on in AI systems is already much better than what's going on in the human brain," Hinton said, arguing for global AI regulation (and adding that it will never happen).
He did not sign the open letter circulated a few weeks ago by the Future of Life organization, which called for a 6-month pause in AI development. Neither had another letter from AI scientists* warning for development. He had not wanted to do that while he was still employed by Google, he said. But now is the time to warn against it, he said. He had also spoken to Alphabet CEO Sundar Pinchai personally about the issue before he left, Hinton said. However, he did not reveal to the NYT what the latter had to say about it.
It is known that Google accelerated the launch of its ChatGPT competitor Bard despite the concerns of its employees. This is what 18 former and current employees told us. Internally, they called the chatbot a "pathological liar" and "embarrassing".