Two out of three Americans believe the government has taken insufficient action to regulate artificial intelligence. This statistic isn’t particularly surprising. However, what stands out in this current political context is the demographic of those Americans: Democrats, independents, and Republicans appear to share this viewpoint nearly equally. On virtually every other topic one could mention, the partisan divide ranges from 65 to 76 percentage points. When it comes to AI regulation, that divide nearly disappears.
This insight emerges from a newly conducted nationwide survey by the Annenberg Public Policy Center at the University of Pennsylvania, which involved 1,330 adult citizens during February and March of this year. The researchers focused on the outlook for the future, and according to most Americans, it appears rather bleak.
A Technology That Users Are Skeptical Of
Almost 80% of participants reported they had heard a significant amount about AI. Around two-thirds indicated they had utilized it, in some capacity, at least a few times in the prior month. These individuals are not unfamiliar with the technology. Yet, when queried about AI’s overall effects on the United States in the next decade, only 17% deemed it positive. Forty-two percent regarded it as negative. A noteworthy 32% expressed a balanced view, considering both positive and negative aspects, reflecting a specific type of ambivalence: not ignorance or indifference, but a clear-eyed apprehension toward something that may have mixed implications.
The only area where genuine optimism surfaced was in healthcare. More than half of those polled anticipated AI would positively influence medical research and advancements, which aligns with intuitive expectations. AI-supported drug discovery, quicker diagnostic imaging, and the potential for genomic pattern recognition at an unprecedented scale are compelling narratives that have emerged. However, when shifting from medicine to nearly any other field, optimism declines sharply. Only 22% believed AI would enhance the creative arts. Only 19% were hopeful for the economy. For mental health and wellness, the figure dropped to 17%. Regarding household utility expenses, possibly influenced by consumers already witnessing rising electricity costs as data centres expand, only 14% had a favorable outlook. The most disheartening statistic: just 5% believed AI would improve US-China relations.
“These findings convey a distinct message,” stated Shawn Patterson Jr., a research analyst at APPC. “Americans are attentive to AI, and their observations raise concerns. The call for regulation transcends party lines. Majorities across the political spectrum indicate the government has fallen short.”
The Most Unusual Bipartisan Agreement in US Politics
To fully understand the peculiarity of the AI data, it is instructive to compare it with other topics examined in the survey. For instance, the partisan divide regarding whether Kamala Harris would have outperformed Donald Trump was 75 percentage points concerning immigration. It was 76 points for inflation. On AI regulation, the difference was 65 points—still significant but the smallest across any tested policy area. More illuminating: 24% believed Harris and Trump would have performed similarly on AI, whereas on immigration, that proportion was merely 7%.
This indicates that AI has not yet been fully appropriated by either political group. It continues to be a contested issue, unaffected by the sorting mechanisms that have shaped nearly every other discussion in American public life. Matt Levendusky, a political scientist at Penn, described it as “noteworthy,” adding, “concern about AI transcends party lines, and the public is eager to see how politicians respond. This presents genuine opportunities for either party to sway public opinion if they can demonstrate that they possess the right solutions.”
In other words, a window of opportunity exists. Whether anyone capitalizes on it remains uncertain.
Even among those who anticipate a very positive outcome from AI’s impact on the country, 43% still believe the government has done too little to manage it. The call for regulation continues even amid optimism.
Data Centres and Local-Level Anxiety
One of the survey’s more tangible inquiries concerned local developments: would you endorse the establishment of new AI data centres in your locality? Almost half of all respondents indicated they would not. Thirty-one percent were strongly opposed, while only 21% supported it. These concerns are not merely abstract notions related to superintelligence or algorithmic bias. They encompass practical issues regarding land use, energy demands, water consumption, noise, and a growing sense that something significant and consequential is being erected in the vicinity without proper community consultation.
Among American workers, 41% expressed some level of concern about losing employment or having their hours reduced due to AI. Democrats exhibited greater anxiety on this matter (50%) compared to Republicans (32%), which may not be entirely surprising given the various industries and jobs that tend to align with each party. However, the anxiety is pervasive. It is,