Edited by Paul Derrington, Chief Editor
TL;DR
Chapman University's 12th annual Survey of American Fears, released October 20, places fear of AI harming the political process at No. 3 (51.9%) and AI hurting the economy at No. 7 (49.7%). The political-process and economy questions were new in 2026, so there is no earlier figure to compare them with; the one AI item Chapman asked before, AI replacing people in the workforce, fell from 49.2% in 2025 to 42.0%. In the April survey, weekly AI users reported less political-process fear than never-users; the association does not establish cause and effect.
The headline numbers
Chapman University released its 12th annual Survey of American Fears on Tuesday, October 20, adding new questions on whether AI could harm the political process or hurt the economy. (Chapman has asked about AI replacing workers since at least 2024, and about AI and robots in 2015.) Two AI concerns landed in the top 10: fear of AI harming the political process ranked No. 3, with 51.9% of respondents afraid or very afraid, and fear of AI hurting the economy ranked No. 7, at 49.7%.
The usual suspects still lead the chart. Fear of corrupt government officials remained No. 1, at 61.1%, and fear of the United States becoming involved in another world war rose from No. 6 in 2025 to No. 2 this year, at 58%. Chapman researcher Christopher Bader said that higher ranking — plus biological warfare and terrorist attacks entering the top 10 — could reflect concerns about the conflict with Iran.
Systemic concerns and job fears
The results extend beyond employment concerns. Chapman's ranking table places AI's potential political and economic harms above cyberterrorism, climate change, and not having enough money for the future. But workforce replacement was also a substantial concern: 42% reported being afraid or very afraid of AI replacing people at work. The survey measures reported fears, not evidence that these harms have occurred.
Reading further down the survey, AI anxiety runs wide. Fear of AI hurting the environment (No. 19) sits just behind fear of high medical bills (No. 18). Fear of AI replacing humans at work and fear of AI harming personal relationships both land in the top 25, just behind climate change and Iran using nuclear weapons.
AI use and political-process fear
Chapman reports that 46% of people using AI at least weekly feared it could harm the political process, compared with 55% of those who never use it. This is a comparison for that specific fear, not a measure showing that users worry less about every AI risk. The association does not establish whether use changes attitudes, existing attitudes influence use, or other factors explain the difference.
Views of emerging technology also split by use case. Respondents were more hopeful than fearful about abundant clean energy, life-extension medicine, and humans working with robots in manufacturing. Continuous employee monitoring, genetically engineering human babies, and AI romantic partners generated more fear than hope. Chapman researcher Christopher Bader interprets the findings as concern about political and economic consequences as well as jobs.
How the survey works
The survey is a project of Chapman's Wilkinson College of Arts, Humanities, and Social Sciences. NORC at the University of Chicago surveyed 1,221 adults online in English from April 7 to April 21, 2026, using its probability-based AmeriSpeak panel. The methods report lists a margin of error of plus or minus 3.79 percentage points for the full sample; that figure should not be assumed to apply to the AI-use subgroups or differences between individual fear rankings.
The fieldwork is a spring snapshot, not a measurement of public opinion on the October publication date. Adding new questions can also change which topics appear in the top 10 without establishing a year-over-year increase in those fears. Concern for loved ones remained prominent: 51.8% reported fear of someone they love becoming seriously ill, and 50.6% feared a loved one's death. Rankings describe reported prevalence; they do not establish statistically meaningful differences between closely placed items.