Utah and National Polls Reveal Diverging Priorities on Artificial Intelligence Risks

Public opinion on artificial intelligence is split, with specific concerns varying by geography and demographic focus.

New polling conducted by the Deseret News in partnership with the University of Utah’s Hinckley Institute of Politics shows that general sentiment toward AI is evenly divided between favorable and unfavorable ratings. The survey then examined the specific drivers behind these feelings. Utah residents ranked privacy and the misuse of personal data as their primary worry. The spread of disinformation and deepfakes followed, with job losses completing the top three concerns for the state's population. The national sample presented a different hierarchy. Loss of human control over AI topped the list of nationwide concerns. Job losses came second, while privacy issues and disinformation tied for third place. Both groups also expressed concern over the loss of human creativity, skills, and the environmental costs associated with energy and water usage for data centers.

Experts reviewed the data to contextualize these findings. Christopher Honey, an AI researcher and professor at Johns Hopkins University, observed a convergence between this poll and his own national studies. He noted that people generally want safeguards for AI while simultaneously hoping for meaningful technological advances. Honey pointed out that privacy concerns were particularly acute in the Utah data. He stated that while privacy showed up in his own data, it was not as high as in the Deseret News results. He added that Utah respondents appeared more concerned about privacy than the general U.S. voting population.

Kenneth Marino, an AI researcher and professor at the University of Utah, attributed the widespread worry to the rapid pace of AI development. He noted that the technology has quickly become ubiquitous in daily life, affecting education, work, and the physical environment through data center construction. He remarked that it is difficult to understate how fast the public emergence of this technology has been.

When asked about sources of hope, both the Utah and national samples aligned closely. The top three hopeful outcomes for both groups involved medical and scientific progress. These included faster medical research and new treatments, earlier and more accurate disease diagnosis, and scientific discovery beyond health. The next three ranked hopes were improving accessibility for people with disabilities, lowering the cost of goods and services, and automating dangerous or repetitive work.

Honey highlighted that previous sentiment surveys had largely ignored these hopes. He described the desire for scientific advancement as a reasonable aspiration. He explained that it is difficult for the public to reason about AI due to its technical complexity, and even developers face substantial uncertainty about the technology's trajectory. Users often do not know if a system remembers previous interactions. Marino, who works on AI for web accessibility, agreed that the field moves quickly. He said that even experts have trouble keeping up. He encouraged the public to voice their concerns to elected representatives rather than acting as passive observers. He emphasized that one does not need to be an expert for those concerns to be valid.

The poll results stand as a snapshot of public sentiment at a time when AI tools are rapidly integrating into everyday workflows and scientific research.