- Three AI godfathers have been arguing about what’s next for the tech they created.
- One of them, Geoffrey Hinton, has been pondering whether the tech should even exist at all.
- However, overemphasizing the dangers of AI risks paralyzing debate at a pivotal moment.
He may be an AI godfather who helped develop the technology that powers an array of chatbots, but Geoffrey Hinton can’t help but ponder if we’d be better off if the technology never existed.
His fears were laid bare in a recent New Yorker profile, where he answered a question about AI in general by saying, “It’s not unreasonable to say, We’d be better off without this — it’s not worth it.”
He added, “Just as we might have been better off without fossil fuels. We’d have been far more primitive, but it may not have been worth the risk.”
While Hinton’s fears are almost certainly genuine, the godlike qualities that AI doomsayers project onto the technology also risk overemphasizing its significance at a pivotal moment.
The language around extinction fears is vague and can sometimes mimic aspects of a theological or existential debate — both distractingly high-brow conversations as governments worldwide inch closer to regulating the tech.
Those who espouse AI’s extinction risks have also faced criticism for shifting the focus away from more urgent threats posed by the technology, paralyzing constructive debate, and allowing Big Tech to control the AI industry.
Hinton has expressed similar concerns before and quit his job at Google to speak openly about the risks of the technology he helped create. But not all his fellow godfathers agree with him.
Yan LeCun, who won the Turing Prize alongside Hinton in 2018, also appears briefly in the New Yorker profile of Hinton.
“I’m not scared of A.I.,” LeCun told the magazine. “I think it will be relatively easy to design them so that their objectives will align with ours.”
The two have publically blasted each other on social media recently, with LeCun sharing a tweet that mocked AI alarmists and called Hinton an “Ultra-leftist, world-class eccentric.”
Creating a fear of the unknown
AI has a “black box” problem, meaning it’s one of the first technologies that even its inventors don’t fully understand.
While Hinton and Meta’s chief AI scientist LeCun have butted heads, fellow collaborator and third AI godfather Yoshua Bengio has stressed that this unknown is the real issue.
“Yann and others are making claims that this is not a problem, but he doesn’t know — nobody knows,” Yoshua Bengio, a Canadian computer scientist and deep learning pioneer, previously told Business Insider. “I think it’s dangerous to make these claims without any strong evidence that it can’t happen.”
For Hinton, the risks AI poses to society — essentially the doomsday threat of control-seeking machines staging a coup against humanity — are pressing and ultimately not ones that he knows how to solve.
“There’s a very general subgoal that helps with almost all goals: get more control,” he told The New Yorker’s Joshua Rothman. “The research question is: how do you prevent them from ever wanting to take control? And nobody knows the answer.”
Hinton is none the wiser about how to mitigate the issues he foresees either. Unlike climate change, he doesn’t see a clear scientific solution.
“There, you know what the solution looks like. Here, it’s not like that,” he said.
Read the full article here