The Majority of Americans Are Now Terrified of AI

A recent Politico poll has revealed that a majority of Americans harbor fears about the potential dangers of artificial intelligence. Nearly two thirds of the public believes there is at least a moderate chance that advanced AI could destroy humanity. This unsettling view is not exclusive to any one political group, as it is prevalent among voters of both Kamala Harris and Donald Trump.

In fact, 70 percent of Harris voters and 60 percent of Trump voters view AI as an existential risk. A significant portion of the population, 17 percent overall, is almost certain that AI could wipe out humanity. This sentiment is evenly split between Harris and Trump voters. However, a sizable majority, 22 percent, still believes that AI poses a very small or no risk.

The poll, conducted in September, reflects the growing alarm surrounding AI as industry leaders have made alarming claims about the risks of their systems' capabilities. Last week, an Anthropic researcher resigned from the company, accusing them of gambling with humanity's lives and claiming that the developers earnestly believe the technology could kill us all.

Anthropic's CEO, Dario Amodei, subsequently called for the US's top AI companies to cooperate on slowing down AI development, garnering support from Elon Musk and OpenAI CEO Sam Altman. Despite some skepticism about the sincerity of AI industry safety claims, the majority of Americans seem to accept the warnings. A substantial 48 percent of Americans agree that AI has become advanced enough that development should be paused outright.

However, there is significant divergence across party lines, with only 44 percent of Trump voters agreeing, compared to 58 percent of Harris voters. Conversely, 28 percent of Harris voters argued for continued development due to potential benefits, compared to 40 percent of Trump supporters. While the threat of AI is widely acknowledged, Americans are divided on how to address it.