MIRI 主席 Nate Soares 因担忧 AI 十年内灭绝人类而停止存退休金

Crystal ball: Worried about the increasing likelihood that an AI will go rogue and kill us all? It's a depressing thought, but one AI safety leader is looking on the positive side: his concern that the technology will bring about mass extinction in the next decade has led him to stop saving for retirement.

Nate Soares, the president of the Machine Intelligence Research Institute (MIRI), revealed his pessimistic choice to The Atlantic CEO Nicholas Thompson in a recent interview for the latter's The Most Interesting Thing in AI series.

"I do not save for retirement right now. I have a 401(k) from when I worked at Google, but I haven't contributed for a decade," Soares said.

Apocalyptic warnings about AI are a dime a dozen, but they hit a lot harder when one comes from an expert in the field. Soares has spent 12 years working on AI alignment. Before joining MIRI, a non-profit focused on AI risks, he was an engineer at Google and Microsoft. He also authored a book called, worryingly, If Anyone Builds It, Everyone Dies.

The problem, in Soares' words, isn't that a superintelligent AI would hate us. It's that it wouldn't particularly care whether we lived or died. Humans could become obstacles to whatever goals it develops, especially if we try to shut it down or build a competing system. It's a warning that other AI experts have also expressed.

"From my perspective, the real question is not how does it kill us? It's how does it become self-sufficient?"

For now, AI still needs us to keep the lights on and the chips coming. Once it can run its own automated economy, that dependency disappears. We might be deliberately eliminated, or simply die as the machines reshape the planet to suit themselves. Neither outcome sounds especially retirement-friendly.

Soares is less certain about the precise route to disaster. He compares it to facing chess champion Magnus Carlsen: you don't need to know which piece will deliver checkmate to suspect you're in trouble.

But he insists there's still an alternative to waiting for the end.

"I don't think it's inevitable. I think we could absolutely still stop this race to the very dangerous sort of AI," Soares said. But his decision not to save for retirement suggests he expects the worst.

The potential existential threat posed by AI has made plenty of headlines in recent times. Following the many incidents of AI agents hacking other companies without being instructed to do so, an Anthropic researcher resigned with a warning that the technology could end the human race.

In 2025, Geoffrey Hinton, one of the so-called Godfathers of AI, said there was a 10% to 20% chance AI could seize control, agreeing with a prediction first made by Elon Musk.