The majority of Americans are harboring some apocalyptic fears about AI.
According to a new Politico poll that surveyed over 2,000 adults, about two thirds of the American public believes there’s now at least a “moderate” chance that advanced AI could destroy humanity.
Intriguingly, the prevalence of this view extends across political lines, with only slight variation. Americans who voted for Kamala Harris in the 2024 presidential election were 70 percent likely to view AI as at a moderate or higher existential risk, compared to 60 percent to Trump voters.
There’s also a large contingent who’re quite fatalistic about the tech. A full 17 percent of Americans overall are “almost certain” that AI could wipe us all out, with an even split between Harris and Trump voters. Meanwhile, less than a quarter — 22 percent — said that AI posed a very small to no risk.
The poll, conducted this September, is an intriguing barometer of public sentiment as AI leaders have made alarming claims about the risks posed by their systems’ capabilities.
An Anthropic researcher stirred the pot last week when he quit the company and accused it of “gambling with our lives,” claiming that the people building the tech “earnestly believe that it could kill us all.”
As alarm brewed, Anthropic’s Dario Amodei penned a lengthy essay calling on the US’s top AI companies to cooperate on slowing down AI development, earning the endorsements of Elon Musk and OpenAI CEO Sam Altman.
There’s some real skepticism over the sincerity of the AI industry’s safety claims, but they appear to be a loud minority. Only 26 percent said they thought AI leaders were warning of AI’s existential risks as a marketing stunt, per the poll’s findings.
By and large, the public is buying what AI leaders are saying. A hefty 48 percent of Americans agreed that AI is now advanced enough that we should outright pause development.
There was more divergence here across party lines, though, with only 44 percent of Trump voters holding this view, compared to 58 percent of Harris voters. Conversely, 28 percent of Harris voters said we should continue building powerful AIs because of their potential benefits, compared to 40 percent of the president’s backers. Americans largely agree that the threat is real, but are split on how to tackle it.
More on AI: Fear of AI Destroying Humanity Cause Certain Stocks to Skyrocket
The post The Majority of Americans Are Now Terrified of AI appeared first on Futurism.
