Dr. Geoffrey Hinton is an AI pioneer, but according to a new interview, he's beginning to feel a bit like Victor Frankenstein studying his monstrous creation. "It is hard to see," Hinton tells the New York Times, "how you can prevent the bad actors from using it for bad things." Hinton, who 11 years ago collaborated with a pair of graduate assistants at the University of Toronto to make the tech that underlies many AI systems today, tells the Times he's quit a position he held at Google in order to speak out about the growing AI threat. Hinton says he regrets what he's done. "I console myself with the normal excuse: If I hadn't done it, somebody else would have."
The Times calls Hinton's flip on AI "remarkable," coming as it does "at perhaps its most important inflection point in decades." Gizmodo notes that even a few months ago, he was referring to AI as a "supernaturally precocious child" and "humanity's butterfly." Hinton is perhaps best known for his development of the backpropagation algorithm—a key technique used to train neural networks. Although he has done vital work in the AI field, Hinton tells the Times that its rapid improvement embodied by tools like ChatGPT or Google's Bard is fundamentally dangerous. Regarding AI, Hinton says, "Look at how it was five years ago and how it is now. Take the difference and propagate it forwards. That's scary."
Hinton grew concerned when he says Google stopped behaving responsibly with AI innovations to compete with ChatGPT and Microsoft's Bing. Now he worries about the burgeoning world of fake imagery and text, noting that most people won't be able to easily tell "what is true anymore." He also fears that a world with "killer robots" is suddenly much more real. He says he thought AI being more clever than human brains "was 30 to 50 years or even longer away." However, he says, "Obviously, I no longer think that." (More artificial intelligence stories.)
,
uBlock 
,
and
Ghostery 
)
