In the troubling view of machine learning researcher Nate Soares, AI is more likely than not to wipe out humanity within the next decade.
At least, that’s what Soares said during a recent episode of the Atlantic’s podcast “The Most Interesting Thing in AI.” During the taping, the AI researcher and president of the Machine Intelligence Research Institute even said that he’s so confident in the coming AI apocalypse that he’s stopped contributing to his retirement account.
“I do not save for retirement right now,” Soares explained during the interview. “I have a 401(k) from when I worked at Google, but I haven’t contributed for a decade.”
The AI guru evidently pegs the possibility of AI-driven extinction at over 50 percent over the next ten years, a remarkably bold prediction of a future which Soares is evidently working hard to prevent.
“I do everything I can to stop this eventuality,” he said. “I don’t think it’s inevitable.”
Soares is well-known in the AI research space as a prolific doom-monger — a schtick which is starting to play on a loop. At MIRI, he operates as Elizier Yudkowsky’s right-hand man, pushing alarmist tales which have secured the pair high-profile media appearances, book deals, and direct access to politicians.
As critics have noted, though both Soares and Yudkowsky have spent much time fretting about an AI apocalypse, there’s no consensus that such a future is even possible — let alone probable.
There are plenty of reasons to be pessimistic about AI. The technology has already been used to bomb elementary school students, scam seniors, and drive dozens of vulnerable users to suicide. AI doesn’t need to be superintelligent to do any of that — and if we all focus on the harms technology may cause down the road, it becomes easy to overlook the harms it’s causing in the here and now.
As for saving your pennies for your retirement? Well, it’s like some AI doomers say — it’s probably better to play it safe and contribute to your savings, rather than take the chance at being sorry you didn’t. Look at it this way: if AI kills us all, it won’t really matter anyway.
More on AI safety: Godfathers of AI Say Models Are So Out of Control That They Need to Be Airgapped for the Safety of Us All