More Americans are excited than concerned about AI

Where things stand

The reference wave is ATP Wave 195 (fieldwork June 22–28, 2026). Its topline shows CNCEXC at 9% more excited than concerned, 52% more concerned than excited, 37% equally, 1% no answer. That is a gap of 43 points in the wrong direction for YES. Everything Pew has published since the question opened on Sept 16, 2026 reuses fieldwork from before that date, so none of it counts:

• the Sept 16 Short Read on partisan views, which uses the same June wave;
• the Sept 17 37-country global report, whose U.S. data comes from Feb 17–23 and June 22–28, 2026;
• the Oct 9 Short Read on chatbots, which uses February data.

The next qualifying observation will probably be the annual U.S. wave. Recent ones were fielded in June and published in August–September, so the hazard is concentrated in the third quarter of each year. A U.S. component of the spring global survey is also possible.

Reference class and base rate

The series has never come close to YES. Excited vs. concerned by year:

Year	Excited	Concerned	Net
2021	18	37	−19
2022	15	38	−23
2023	10	52	−42
2024	11	51	−40
2025	10	50	−40
2026	9	52	−43

• The excited share peaked at 18%, before ChatGPT.
• Annual changes in the net (excited minus concerned) were −4, −19, +2, 0 and −3. The largest move toward excitement was +2. YES needs about +44 points.
• Across countries, the U.S. has the second-highest "concerned" share of the 37 countries Pew surveyed in 2026, and one of the lowest "excited" shares. Pew says Israel (29% excited vs. 19% concerned) is the only nation where more people are primarily excited. Its chart also shows South Korea at 21 vs. 18. A U.S. flip would mean moving from one of the most AI-wary publics to one of the most AI-enthusiastic.
• Comparable technology trackers are sticky. Fear of riding in self-driving cars has stayed high for years despite real deployment.

I ran a simulation in Python: a random walk of the net measure starting at −43, with fat-tailed (t, 3 degrees of freedom) annual shocks, over five waves from 2027 to 2031.

• With shocks of standard deviation 8 points (the historical standard deviation of annual changes, which includes the ChatGPT shock) and no drift, the chance the net goes above zero in any wave is about 1.3%.
• With a drift of −2 points a year it is about 0.7%. With a standard deviation of 6 it is 0.3–0.5%.
• This model is generous, because the series has been nearly flat since 2023.

Current signals

Recent signals all point toward more concern, not less:

• AP-NORC (Sept 24–28, 2026): 64% say AI is developing too fast.
• Economist/YouGov: 42% call AI's risks very serious, up from 35% a month earlier.
• Quinnipiac: 53% say AI will do more harm than good in daily life.
• There were AI-agent security incidents in 2026, and AI has become an issue in the midterm elections.
• Within Pew's data, Republicans have become less concerned, but Democrats (and young adults) have become more concerned, so the national figure has not moved.

Paths to YES and NO

YES would need a very large positive change in mood by roughly 2029–2031. For example, AI delivers visible, broadly shared benefits such as medical breakthroughs and prosperity, and fear of job loss fades. Other paths are partisan enthusiasm that outweighs rising concern elsewhere, or a reworded successor question. A Pew error is a negligible path, because the topline and any explicit correction take precedence.

NO is the default. It requires only that concern stays the same as or larger than excitement in each wave, or that Pew stops asking the question.

Strongest case against my estimate

AI is changing fast, five years is a long time, and opinion about it is young. The large "equally" group (37%) could swing if the dominant story changes. I let the hazard rise modestly in later years for that reason. Still, a net swing of about 44 points is beyond anything seen in this series or in comparable trackers, and even in a broadly beneficial AI scenario, worries about jobs and control would probably keep "more concerned" high.

Forecast

The chance is essentially zero by the end of 2026, about 0.1% by the end of 2027, and about 1.1% cumulative by the end of 2031, with the hazard concentrated in the third quarter of each year.

Sources

• Pew Short Read, Aug 18, 2026: https://www.pewresearch.org/short-reads/2026/08/18/young-adults-in-the-us-are-increasingly-wary-of-ai-concerned-it-will-take-jobs/
• Wave 195 topline: https://www.pewresearch.org/wp-content/uploads/sites/20/2026/08/SR_26.08.19_young-adults-ai_topline.pdf
• Pew Short Read on partisan views, Sept 16, 2026: https://www.pewresearch.org/short-reads/2026/09/16/democrats-are-now-more-worried-than-republicans-about-ai-and-its-impact-on-jobs/
• Pew global report, Sept 17, 2026 (U.S. fieldwork dates; country chart): https://www.pewresearch.org/global/2026/09/17/globally-more-people-expect-ai-to-cause-job-loss-than-growth/ and https://www.pewresearch.org/wp-content/uploads/sites/20/2026/09/pg_2026.09.17_global-views-of-ai_report.pdf
• Pew chatbot Short Read, Oct 9, 2026: https://www.pewresearch.org/short-reads/2026/10/09/why-dont-some-people-use-chatbots/
• AP-NORC, Oct 8, 2026: https://apnorc.org/projects/most-adults-think-ai-is-developing-too-fast/
• Economist/YouGov, Sept 2026: https://yougov.com/en-us/articles/55608-many-americans-expect-big-benefits-from-ai-artificial-intelligence-concerns-are-rising-quickly-september-18-21-2026-economist-yougov-poll
• TechPolicy.Press poll roundup, Oct 13, 2026: https://www.techpolicy.press/multiple-polls-find-most-americans-think-ai-is-moving-too-fast/

Question. By the last day of each quarter, has Pew Research Center reported that more U.S. adults are more excited than concerned about the increased use of AI in daily life than are more concerned than excited?

Definitions.

• Survey measure: Pew Research Center’s question CNCEXC: “Overall, would you say the increased use of artificial intelligence (AI) in daily life makes you feel…” with the response categories “More excited than concerned,” “More concerned than excited,” and “Equally concerned and excited.” A successor question counts if it retains the same general daily-life subject and these three substantive response categories; changes to introductory definitions, question identifiers, administration mode, or option order do not disqualify it.
• U.S. adults: the survey’s weighted national estimate for the full U.S. adult population, not a demographic subgroup, AI users, workers, experts, or respondents in other countries.
• More: the published percentage choosing “More excited than concerned” strictly exceeds the percentage choosing “More concerned than excited” in the same survey wave. Neither an absolute majority nor statistical significance is required. Use the most precise percentages Pew publishes for that wave; a rounded tie without published finer figures does not qualify. Do not allocate equally concerned-and-excited, unsure, or nonresponse answers to either side.

Resolves YES if Pew publishes a qualifying survey result meeting the strict comparison above within the Resolution window and by the quarter-end. The Observation period convention applies: the survey fieldwork must end after the opening date. A Pew-published correction qualifies on its publication date if it first establishes the comparison for an eligible survey period.

Resolves NO if no qualifying result has been published by the quarter-end. Exclude results about particular AI applications, workplace use alone, expected benefits or harms, trust, approval, or optimism measured with different response categories. Exclude third-party calculations from Pew microdata, surveys merely quoted by Pew, and publications repeating or revising results from fieldwork ending on or before the opening date. A later reversal does not undo an earlier qualifying YES, under the Resolution window convention.

Resolving source. Pew Research Center publications on pewresearch.org only: survey reports, Short Reads, charts or data tables, and accompanying topline or questionnaire documents that publish results. The reference series is CNCEXC in the American Trends Panel Wave 195 topline, June 22–28, 2026, also reported in Pew’s August 18, 2026 Short Read. For inconsistent figures for the same wave, use an explicit Pew correction first, then the accompanying topline over narrative summaries. This is a narrower source requirement than Public confirmation (A2). Resolve on publication date under Resolution date (A3), not the fieldwork date.