PHILADELPHIA — Public resistance to new data centers has risen sharply across the United States, even as Americans’ overall views of artificial intelligence remain largely unchanged, according to a nationally representative survey from the Annenberg Public Policy Center at the University of Pennsylvania. The findings reveal a growing divide between public attitudes toward AI as a technology and attitudes toward the physical infrastructure required to operate it. While many people continue to see potential benefits from AI, particularly in medical research, opposition increases substantially when the technology is associated with a large industrial facility being built near their homes.

The survey found that 61% of U.S. adults somewhat or strongly oppose the construction of new data centers in their communities. That figure is 12 percentage points higher than the 49% recorded in an earlier APPC survey conducted in February and March. Support fell during the same period, from 21% to 14%. The change comes as communities across the country debate the environmental, economic, and social consequences of rapidly expanding computing infrastructure used to train and operate advanced AI systems.

Data centers are specialized facilities filled with servers, networking equipment, cooling systems, backup power supplies, and security infrastructure. AI applications can require enormous amounts of computational power because they process vast datasets and perform calculations across billions of model parameters. Training a large AI model may require thousands of processors operating simultaneously, while everyday use also depends on data centers to generate responses, analyze images, translate languages, and carry out other tasks. These facilities can consume substantial amounts of electricity and water, making their construction a highly visible local issue even when the software they support is used remotely.

Opposition to local data centers crosses political boundaries. Majorities of Democrats, Republicans, and independents said they opposed new facilities in their areas, although the proportions differed: 69% of Democrats, 54% of Republicans, and 53% of independents expressed opposition. Resistance was strongest among adults younger than 30, reaching 70%, and declined to 57% among people 65 and older. The pattern is notable because younger adults are often assumed to be more comfortable with emerging technologies and are more likely to use AI-powered services in education, work, entertainment, and communication.

Despite the shift in views about local infrastructure, broader opinions about AI have remained stable. Thirty-nine percent of respondents said they expected AI’s impact on the United States over the next decade to be somewhat or very negative, compared with 18% who expected a positive impact. The result was statistically unchanged from the spring survey. In addition, 68% said the federal government has done “too little” to regulate AI, a view shared by majorities of Democrats, Republicans, and independents. The results suggest that concerns about AI regulation have become persistent rather than being driven solely by short-term news events.

Respondents distinguished sharply between areas where AI might produce public benefits and areas where it could create risks. Among 13 sectors examined, medical research and scientific discovery was the only category receiving a net-positive assessment, with anticipated benefits exceeding expected harms by 41 percentage points. AI systems are increasingly being used to identify patterns in biological data, predict molecular structures, assist with medical imaging, and help researchers screen potential treatments. These applications may accelerate work that would otherwise require years of laboratory experiments, although they still depend on human validation and carefully controlled clinical studies.

The most negative assessments involved personal privacy and data security, which registered a net score of minus 63 points. Children’s online safety followed at minus 50 points, while employment and jobs reached minus 46 points. These concerns reflect the technical and social consequences of deploying AI at scale. Models may be trained on or exposed to sensitive information, automated systems can make decisions that are difficult to explain, and workplace applications may alter or eliminate certain tasks. The survey does not measure whether these risks will materialize in specific industries, but it indicates that many Americans view them as more immediate than AI’s possible benefits outside medicine.

Personal experience with AI was associated with more favorable opinions about its national impact. Among respondents who said they had never used AI during the previous month, 54% expected its effect on the country to be negative. That proportion fell to 40% among light users and to 29% among those who used AI many times or almost every day. However, frequent use did not make people more supportive of data centers. Opposition to building new facilities near respondents’ homes remained essentially flat across different levels of AI use. This distinction suggests that familiarity may influence perceptions of digital tools, while concerns about electricity demand, water consumption, noise, traffic, land use, and local utility costs may remain regardless of personal experience with AI software.

“The people who use AI the most are the most optimistic about what it will do for the country, and that has been consistent across our surveys,” said Matt Levendusky, a University of Pennsylvania political science professor and the Stephen and Mary Baran Chair in the Institutions of Democracy at APPC. He said that optimism had clear limits, particularly when questions involved privacy, employment, or the construction of a data center nearby. The survey was conducted by SSRS for APPC’s Institutions of Democracy division among 1,320 U.S. adult citizens from June 16 through July 19, 2026. Respondents were surveyed primarily online, with a small telephone sample, and weighted to match population benchmarks. The margin of error for the full sample was plus or minus 3.5 percentage points, with larger margins for subgroups. Together, the findings indicate that the next phase of the AI debate may be shaped less by abstract promises about intelligent software than by negotiations over the highly physical systems that make it possible.

Subject of Research: People

Article Title: Opposition to Local Data Centers Rises Sharply as Views on AI Hold Steady

Web References: Annenberg Public Policy Center news release; Reuters; The New York Times; Associated Press

References: Annenberg Public Policy Center Institutions of Democracy national survey, June 16–July 19, 2026; APPC AI topline survey data; APPC survey methodology

Image Credits: Annenberg Public Policy Center

Keywords: Artificial intelligence, data centers, public opinion, AI regulation, medical research, privacy, data security, employment, social surveys, political science

A new study is drawing attention to a hidden problem in mental-health research: scientists may be using the same words to describe first episode psychosis (FEP) while measuring entirely different things. In an analysis published in Schizophrenia, researchers Seiber, Sridhar, Hasenstab and colleagues examine how U.S. insurance claims data are used to estimate the incidence of FEP—the number of new cases appearing in a population over a defined period—and argue that greater standardization is urgently needed.

First episode psychosis refers to the point at which a person experiences psychotic symptoms for the first time in their life. These symptoms can include hallucinations, delusions, disorganized thinking or major changes in behavior and functioning. FEP is not a single diagnosis. It can occur in schizophrenia-spectrum disorders, bipolar disorder, severe depression and several medical or substance-related conditions. Because early treatment is linked to better long-term outcomes, knowing how often FEP occurs is essential for planning specialized clinics, staffing early-intervention programs and directing public-health resources.

The researchers focus on U.S. claims data, enormous digital records generated when patients receive medical care and providers submit bills to insurers. These databases can include diagnostic codes, procedure codes, prescription information, dates of service and details about emergency, inpatient and outpatient treatment. Unlike traditional epidemiological surveys, claims datasets can cover millions of people and make it possible to study patterns across large geographic areas. Their scale, however, does not automatically make their estimates precise.

The central challenge is that a billing code is not the same as a clinical assessment. A code for psychosis may indicate a confirmed diagnosis, a suspected condition, a historical problem or a provisional label assigned during an emergency evaluation. Claims data also rarely contain the complete clinical narrative needed to determine whether symptoms truly represent a person’s first psychotic episode. A patient may have received care in another health system, paid out of pocket, changed insurers or carried an earlier diagnosis that is invisible in the database being studied.

Even the definition of “incidence” can vary. Some researchers count a first-ever psychosis-related diagnosis, while others count a first hospitalization, a first emergency-department visit or the first appearance of a relevant code after a period without documented care. These approaches can produce different numbers from the same underlying population. A study requiring a long “clean period”—a span of time with no previous psychosis-related claims—may reduce the likelihood of counting established cases as new, but it can also exclude people whose earlier treatment occurred outside the available records.

Age, sex, insurance type and geography can further influence the apparent rate of FEP. Young adults may be more likely to encounter emergency services, while people in rural areas may face limited access to psychiatrists and specialized programs. Differences in insurance coverage can determine which services generate observable claims. If researchers compare estimates without accounting for these factors, they may mistake variations in detection or access for genuine differences in disease occurrence.

The study’s call for standardization points toward a more consistent technical framework. Researchers need to specify which diagnostic codes qualify as psychosis, whether substance-induced and medically caused episodes are included, how long a person must remain free of prior claims to be considered a new case, and which healthcare settings are examined. They also need to define the population denominator—the number of people considered at risk—and explain how individuals with incomplete enrollment or interrupted insurance coverage are handled.

Validation is another crucial step. Claims-based algorithms should be compared with detailed chart reviews, clinical registries or structured assessments to determine how accurately they identify genuine FEP. Two statistical measures are especially important: sensitivity, or the ability to capture true cases, and positive predictive value, or the proportion of algorithm-identified cases that are confirmed after review. An algorithm with high sensitivity may find more possible cases but also produce more false positives; one with high specificity may miss people who received vague or inconsistent coding.

Standardized methods could make FEP estimates more comparable across states, insurers and research teams. That would help scientists detect real trends rather than methodological noise and could reveal whether early psychosis services are reaching the populations most in need. More reliable estimates could also strengthen efforts to investigate racial, socioeconomic and regional disparities, provided that claims data are interpreted carefully and not treated as a complete substitute for clinical information.

The message from the research is not that U.S. claims databases are unusable. On the contrary, their breadth makes them one of the most powerful tools available for studying mental-health care at scale. But their value depends on transparent definitions, validated case-finding methods and clear reporting standards. As health systems increasingly rely on administrative data and automated analytics, agreeing on what counts as a new case of psychosis may be the difference between a map that reveals a public-health crisis and one that merely reflects the quirks of the billing system.

Subject of Research: Estimating first episode psychosis incidence rates using U.S. claims data

Article Title: Estimating first episode psychosis (FEP) incidence rates using U.S. claims data: a need for standardization

Article References: Seiber, E.E., Sridhar, S., Hasenstab, K.A. et al. “Estimating first episode psychosis (FEP) incidence rates using U.S. claims data: a need for standardization.” Schizophrenia (2026). https://doi.org/10.1038/s41537-026-00793-4

Image Credits: AI Generated

DOI: 10.1038/s41537-026-00793-4

Keywords: First episode psychosis, FEP, incidence rates, U.S. claims data, schizophrenia, mental health, epidemiology, healthcare data, standardization