A new Reuters/Ipsos poll suggests artificial intelligence has become a source of concern for US voters far beyond the question of new technology. A majority of registered voters surveyed said the Trump administration and Congress are not taking AI risks seriously enough, while 84% said they viewed AI as a threat to American workers.
The survey also found that 62% believe AI could get out of control and endanger the future of humanity. At the same time, 69% of voters backed stricter government regulation of AI. The findings come as Washington weighs how to balance rapid development against concerns over employment, safety and the growing footprint of data centres.
The poll was conducted online from 30 September to 5 October among 4,506 US adults, including 3,526 registered voters, with a margin of error of two percentage points for the full sample.
AI Jobs Fears Cut Across US Political Divides
The employment question appears to be one of the strongest sources of anxiety surrounding AI. That figure was higher than the shares identifying illegal immigration, foreign-worker competition or large corporations as threats to US jobs in the same survey.
Daniel Kokotajlo, a former OpenAI employee who left the company in 2024, told Reuters that job losses and threats to health and safety may be more immediate concerns for the public than the possibility of AI eventually becoming uncontrollable.
His comments highlight a distinction between longer-term concerns about AI becoming uncontrollable and more immediate concerns about jobs, health and safety.
Views were also sceptical of the technology companies themselves, with voters expressing concern about whether AI firms were taking the technology's risks seriously.
The concern is not confined to one party. Sixty-one per cent of Republicans and 80% of Democrats supported stricter government regulation of AI. The figures show support for stricter AI regulation across both major parties, although support was higher among Democrats.
Trump AI Policy Faces Regulation and Data Centre Questions
Federal action has so far been limited, with Trump announcing voluntary safety principles agreed to by major technology companies.
On 29 September, Trump announced that major technology companies had agreed to safety principles that include working with independent auditors to examine whether AI systems operate as intended and taking steps to prevent unauthorised access to technical systems.
The agreement is voluntary and does not specify penalties for companies that decline to comply.
Trump has also continued to support the expansion of US data centres, arguing that the infrastructure is important for maintaining American competitiveness in AI. That position is becoming harder to separate from local concerns about electricity demand and utility costs.
The US Energy Information Administration said this week that electricity consumption rose 4% nationally in the third quarter from a year earlier and expects commercial and industrial electricity demand to grow by close to 3% in 2027. The agency says data centres are included in the commercial sector driving part of that demand.
Support for limits on new data-centre construction reached 69% among Democrats, 55% among independents and 43% among Republicans. Those figures suggest that the debate is extending beyond abstract questions about AI safety into disputes over infrastructure and its effect on communities.
The poll also found that 57% of voters said the US government should not use AI to determine the target of a military strike, compared with 49% in a Reuters/Ipsos survey conducted in August 2025. Meanwhile, 55% of Americans said in an earlier Reuters/Ipsos poll that slowing AI development would be good for the United States.
The findings do not, on their own, show that voters want AI development stopped altogether. Instead, the findings show voters weighing questions about how the technology is governed, where its costs fall and what protections exist if its effects reach workplaces, communities or national security.
