Artificial intelligence might be tearing the worlds of government and business lately, but for decades AI research was the realm of obscure researchers who laid the groundwork for the gold rush we’re seeing today.
Now that their baby has gotten big, it’s always interesting watching these figures grapple with the world they’ve created. For a perfect example, look no further than a newly-resurfaced talk in which AI pioneer Geoffrey Hinton gave a fascinating response when asked a question about superintelligent AI replacing humans.
During the December online lecture, which recently resurfaced on Reddit, the so-called “Godfather of AI” was asked if he would be in favor of a superintelligent AI destroying humanity and replacing it with “something objectively better in terms of consciousness.”
“I’m actually for it,” the machine learning scholar responded, “but I think it would be wiser for me to say I am against it.”
Hinton, as you may recall, left Google last spring over fears of AI falling into the hands of “bad actors” and went so far as to say he regretted his role in developing the powerful neural networks that now undergird everything from ChatGPT to Stable Diffusion.
Given that fairly recent turn of events, it’s surprising that he of all people would make such a remark — and when lecture moderator and MIT tech historian Deborah Fitzgerald asked him to “say more,” things did indeed get more and interesting and complicated.
“Well, people don’t like being replaced,” Hinton said laughingly.
At Fitzgerald’s prodding, he expounded by pointing out that there are lots of good and “not-so-good” aspects to humanity, “it’s not clear we’re the best form of intelligence there is.”
“Obviously, from a person’s perspective, everything relates to people,” Hinton continued. “But it may be there comes a point when we see things like ‘humanist’ as racist terms.”
After a long pause, the MIT moderator responded with an “okay” before moving on to the next question. But all the same, this response is all the more perplexing given that just prior to resigning from Google, Hinton told CBS News it’s “not inconceivable” that human-level artificial general intelligence (AGI) could destroy humanity as we know it.
We’ve reached out to Hinton to ask if his views have changed since leaving Google or if he’d like to expound on his remarks.
More on turbo-charged AI: Edward Snowden Says OpenAI Just Performed a “Calculated Betrayal of the Rights of Every Person on Earth”
