takeoff.watch

25% of Americans name AI the country's most important problem

Will Gallup report that at least 25% of US adults name AI as the most important problem facing the country in any monthly poll?

Chance by Q4 2031

12%

Cumulative probability

0%25%50%75%100%
  • Latest forecast

Forecast history

Forecast made on 16 Oct 2026

Latest of 5 forecasts · Methodology v2.1

Model reasoning

Forecast made 16 Oct 2026 · Methodology v2.1

Only about 1 in 100 Americans name technology or AI as the country's top problem in Gallup's monthly survey, even after a summer of alarming AI incidents and warnings from AI company leaders. Reaching 1 in 4 would put AI alongside the biggest shocks in the survey's history, like the September 11 attacks or COVID. I put the chance at about 12% over five years. A major AI-caused disaster, or mass job losses blamed on AI, would most change this.

Where things stand

  • Gallup has no AI category yet. In the most recent Gallup wave (Sept 1–17, 2026; topline published Oct 13, 2026), the only bucket that could hold AI answers is "Advancement of computers/technology." It stood at 1%. Monthly readings for March–September 2026 were *, 1, *, *, , 1, 1 ("" means under 0.5%). From June to November 2025 it was *, and in December 2025 it was 1%.
  • What leads the list: government/poor leadership at 30%, cost of living at 13%, the economy at 11%, immigration at 10% and the Middle East war at 7%.
  • Most of that September wave came before AI news peaked. The wave covered only part of the busiest stretch of AI news so far:
    • the OpenAI–Hugging Face "rogue agent" hack;
    • an Anthropic researcher resigning in public over AI risk;
    • Dario Amodei's Sept 12 essay saying an agent swarm might be able to take over the internet within 6–12 months;
    • lab chief executives endorsing a slowdown;
    • GPT‑6 Astra being rated "Critical" for cyber risk.
  • The October wave is still in the field. Agent incidents and Senate hearings have kept AI in the news, so I expect AI to rise somewhat, but not to anything near 25%.
  • Other polls agree that AI is still low on people's list. Reuters/Ipsos (September 2026, a single choice from a fixed list) had AI at 4% as the top problem. AP‑NORC found 35% of voters rate AI as "one of the most important issues," but that question lists issues for people to rate. Gallup's question is open-ended, and on that format AI is about 1%.

Base rate

  • 25% is a crisis-level reading. Since about 1990, issues that started near zero and reached 25% or more in Gallup's monthly reading have been rare: terrorism after 9/11 (46%), the Iraq war, and COVID (45%). Long-running issues have peaked near this level, for example immigration at a record 28% in February 2024. No technology topic ever has.
  • Annual data shows how rare such jumps are. I used the Comparative Agendas Project's annual Gallup dataset, which runs from 1947 to 2023 and is coded by broad topic (https://minio.la.utexas.edu/compagendas/datasetfiles/US-Public-Gallups_Most_Important_Problem-21.3.csv). In it, about 10 topics jumped from under 5% to 15% or more of annual responses within five years. That works out to roughly a 3% chance per topic in any five-year window.
  • The science/technology/communications topic never got above about 8%, and most of that was mentions of the media.

Adjusting for this case

AI is not a random topic. In this world, capabilities are moving very fast (an AI-generated solution to the Navier–Stokes Millennium Prize problem, labs acknowledging that AI is starting to help build the next AI, autonomous hacking incidents). There are three main routes to YES:

  1. A widely felt AI disaster, such as a swarm or cyberattack disrupting the internet, banks or power. Roughly 15–20% likely by 2031; if it happens, maybe a 30–40% chance that AI reaches 25%.
  2. Large job losses blamed on AI, around 2029–31. Roughly 20% likely; if it happens, about a 35% chance. Many people would answer "jobs" or "the economy" rather than naming AI.
  3. AI becoming the central issue of the 2028 campaign without either of the above. About 3%.

After allowing for overlap between these routes, I get about 13%. I then trimmed for two further risks:

  • Gallup might drop or change the series. It stopped measuring presidential approval in February 2026.
  • Gallup might not report AI as its own category.

Outside forecasts point the same way. Sentinel's forecasters (Sept 14, 2026) put tech/AI reaching 10% before 2028 at 14%. Since 25% is a much harder bar, that is consistent with my roughly 2.6% for 25% by the end of 2027. In the AI‑2027 scenario, an extreme fast-takeoff story, the authors' figure was 20% at the height of a public backlash. Real events that resemble that backlash have so far moved Gallup's number only to 1%.

Hazard profile

Quarterly chances that it first happens, used in the forecast:

  • 0.25% for the fourth quarter of 2026. Only the October–December waves can count, and no disaster has happened.
  • 0.5–0.65% per quarter through 2028. This covers the near-term disaster risk Amodei described and the 2028 election.
  • About 0.6% per quarter in 2029–30, when job losses could peak.
  • 0.55% per quarter in 2031.

Adding these up gives a cumulative 11.6% by the end of 2031.

Strongest case against my estimate

Lower: this measure barely moves. Even the most intense AI news month on record produced 1%, and 25% means a 25-fold increase. Lower still would be reasonable.

Higher: this world looks like the fast-progress scenarios. Superhuman AI that disrupts jobs and politics by 2029–31 could make AI the dominant national problem. I put about half that much weight on this case.

Sources

Resolution criteria

Opened 16 Sep 2026Not yet resolved

Question. By the last day of each quarter, has Gallup reported that at least 25% of U.S. adults name AI as the most important problem facing the country in a monthly poll?

Definitions.

  • Monthly poll: a single monthly national survey wave in Gallup’s U.S. Most Important Problem series, asking the open-ended question “What do you think is the most important problem facing this country today?” or substantively equivalent wording. Fieldwork may span calendar months. A multi-month average does not count.
  • U.S. adults: Gallup’s overall national adult population estimate, not a demographic, partisan, geographic or voter-only subset.
  • Name AI: a response Gallup codes as “artificial intelligence,” “AI,” or an explicitly AI-specific equivalent, including AI-specific harms or concerns. A broader category such as technology, automation, unemployment or data centers counts only if Gallup separately reports the share of all adults whose responses explicitly concern AI. Multiple AI-specific categories count together only when Gallup publishes a combined percentage that counts each respondent once.
  • At least 25%: Gallup’s published point estimate is 25% or higher. A published whole-number 25% counts; a published 24.9% does not. Error bounds do not count. The denominator must be adults, not the total number of problem mentions.
  • Most important problem: any qualifying mention in response to the series’ open-ended question; AI need not be the respondent’s first or only mention, nor the highest-ranked problem nationally. Gallup’s series permits multiple mentions.

Resolves YES if Gallup publishes a qualifying monthly estimate meeting the threshold. Apply the Resolution window and Observation period conventions: the publication must be dated on or after 2026-09-16 and the survey fieldwork period must end after that date. A later Gallup correction or newly published breakdown may qualify on its own publication date if its observation period qualifies.

Resolves NO at a horizon if no qualifying publication exists by that horizon. Prompted lists of problems, AI-specific attitude questions, polls by other organizations, Gallup World Poll results, and estimates inferred from broader response categories do not count. Discontinuation or absence of an AI category does not by itself establish the threshold. Once YES, it remains YES under the Resolution window convention.

Resolving source. Gallup-authored publications only: the Most Important Problem historical-trends page, Gallup News articles, and Gallup’s published survey toplines, tables or downloadable datasets for this series. These are a narrower source requirement than Public confirmation (A2). Use the dated release, dated update or dated Gallup archival publication establishing the figure; apply Resolution date (A3), not the fieldwork date. The series’ open-ended, multiple-response format and monthly cadence are documented in Gallup’s April 30, 2024 report.

Download all forecasts (JSON)