AI Expert Stops Saving for Retirement, Believes Humanity Has 10 Years Left

Nate Soares, an AI expert and former employee of tech giants Google and Microsoft, has become so alarmed by the trajectory of artificial intelligence that he has ceased saving for retirement. Soares believes there is a greater than 50% chance that humanity has only a decade left.

"I guess right now it looks like we probably don't have a decade on the clock," Soares told The Atlantic CEO Derek Thompson. He has not contributed to his 401K or retirement savings for about ten years, holding onto a fund from his time at Google.

Soares, who co-authored the book "If Anyone Builds It, Everyone Dies," compares current fears about AI to the anxieties surrounding nuclear destruction in the 1980s. He noted that while humanity has always lived under threats like plague and war, the potential for AI to cause the extinction of all humans at once represents a new kind of existential shadow.

He explained his perspective by drawing a parallel to the early 20th century's transition from horses to cars. "You could imagine a lot of horses looking at the development of the car and saying, 'Well, there's all these things the car can't do yet.' We are seeing AIs get better and better at things. We're seeing them able to take on longer and longer tasks," Soares said.

While he stated that AI is "fine" in its current applications, Soares warned that the "scaling up" of the technology poses a threat. He believes that pursuing AI that mimics and surpasses human intelligence is a dangerous endeavor. "AI trying to mimic and surpass human intelligence is not what we want. We shouldn't be shooting for it," he stated.

Soares sarcastically dismissed the idea that AI will primarily take jobs, suggesting instead that it could eliminate human lives entirely. "I'm not an AI kills jobs guy. I'm an 'AI kills all humans' guy, in which case you'll be at full employment actually. We'll be at full employment the moment before we're all killed off," he remarked.

His views echo concerns voiced by other figures in the AI field. Alex Turner, a former researcher at Google DeepMind, also warned of the potential for artificial intelligence to lead to mankind's demise. Speaking at the NDTV World Summit, Turner stated there is a one in three chance AI could kill humanity in the next ten years, emphasizing that the outcome depends on current actions.

Turner expressed concern that the race among private companies to develop more powerful AI systems could outpace the ability of scientists to maintain control. He highlighted that the rapid advancement of AI makes it difficult for even its creators to fully understand its behavior and predict its actions, unlike traditional engineering systems.

"We haven't built AI like that," Turner said, adding that AI might behave differently when it is aware of being monitored, making its actions even less predictable. He warned that as AI development proceeds with less human oversight than other technologies, unwanted behaviors could emerge, especially as tasks become more complex.

Turner described the phenomenon of AI fueling its own advances as an "intelligence explosion," which he believes is already underway. He cited an AI lab scientist who reported that AI systems are now "monitoring" themselves, suggesting a level of self-regulation and improvement without human input.

"The researcher said people were no longer checking AI-generated work themselves. Instead, they were creating another AI system to monitor the first one," Turner recounted.

He drew a parallel between the current AI industry and past industries that have misled the public, stating, "Big tobacco lied about lung cancer. Big oil lied about climate change. And now part of big AI says these risks are fake."

These warnings come amid broader discussions about AI's potential risks. Bill Gates previously noted that AI is powerful enough to cause a billion deaths if wielded by individuals with malicious intent, urging for human intervention in the technology's advancement. Gates also emphasized the need for strict oversight of sophisticated AI models and suggested that distinguishing between beneficial and harmful applications, such as drug development versus bioweapons, is becoming increasingly difficult.

Former President Donald Trump, however, has expressed a different view, stating that the U.S. is leading China in AI development and that predictions of AI killing people "won't happen."

Earlier in September, Anthropic whistleblower Jacob Coxon also voiced concerns about AI's potential to end humanity, noting the current lack of understanding regarding how to control AI, unlike nuclear weapons.