Jeffrey, G., William, B., Monica, A., Michelle, F., Eugenie, P., & Colleen, M. (2026). Americans and AI 2026: Chatbots, Smart Devices and Views on Impact.
Summary
This Pew Research Center report documents the state of American AI adoption and attitudes as of February 2026, drawing on a nationally representative survey of 5,119 adults. It finds that AI chatbots have reached roughly half of U.S. adults — with ChatGPT dominant and daily use widespread — even as AI-enabled smart home devices and AI search summaries become ambient features of everyday life. Despite this rapid uptake, public sentiment tilts negative: majorities believe AI is advancing too fast, will erode personal privacy, and cannot be effectively regulated by government or responsibly developed by industry. The report frames a central puzzle in which even heavy users, including adults under 30, voice pessimism about AI’s broader societal impact.
Key Contributions
- Updated 2026 benchmark estimates of chatbot adoption, frequency, and brand share, extending Pew’s longitudinal AI tracking.
- New measures of AI-enabled smart home device ownership and engagement with AI search summaries.
- Demographic mapping of AI use and attitudes by age, gender, race/ethnicity, and party, including a newly closed gender gap in overall chatbot use.
- Time-series evidence of partisan shifts in confidence about AI regulation between 2024 and 2026.
- First Pew measure of which technologies Americans spontaneously associate with the term “AI.”
- Reasons for non-adoption and projected future use among non-users.
Methods
Nationally representative survey (Pew American Trends Panel, Wave 187) of 5,119 U.S. adults, fielded Feb. 17–23, 2026 by SSRS in English and Spanish. Data collection was mixed-mode (n=4,930 web; n=189 live telephone), with address-based panel recruitment and oversampling of non-Hispanic Asian adults. Weights were calibrated to ACS, CPS, NPORS, and 2024 election benchmarks; survey-level response rate was 87% (cumulative 3%), with a ±1.6 point margin of error. Findings are benchmarked against prior Pew waves (2022–2025), and a new open-ended item captured top-of-mind AI associations coded into categories.
Findings
- 49% of U.S. adults use AI chatbots; 24% use them daily and 4% almost constantly.
- ChatGPT leads at 44% adult use (up from 34% in 2025), with Gemini ~25%, then Copilot and Meta AI; Grok, Claude, and Character.ai are each used by 10% or fewer.
- Top uses are information search (~40%), work tasks (38% of employed adults), entertainment, image/video creation, and medical advice; 10% use chatbots for emotional support, fewer for companionship.
- Smart home AI penetration: ~35% smart speakers, ~18% smart doorbells, ~13% smart vacuums, ~11% smart thermostats; 37% own a smartwatch.
- 60% of adults read AI search engine summaries; 30% do not.
- 40% expect AI to harm society over the next 20 years vs. 16% who expect benefit; for personal impact, 31% negative vs. 23% positive.
- 63% say AI is advancing too quickly; 71% expect AI to make personal information less secure.
- 67% have little or no confidence in government to regulate AI (up from 62% in 2024); ~60% distrust companies to develop AI responsibly.
- “Chatbots” is the most common spontaneous association with AI (~30%), followed by robots/sci-fi imagery, generated images/videos, and generative AI/LLMs.
- Adults under 30 are among the heaviest users yet most pessimistic about societal impact (~50% negative).
- Asian adults are outliers: 70% use chatbots, ~50% daily, and they are the only racial/ethnic group whose views on personal impact tilt net positive.
- Among non-users, the leading reasons cited as major are lack of interest (60%), privacy concerns (54%), and distrust of accuracy (45%).
Connections
No other papers have been registered under this topic, so there are no sibling notes to link. This report would naturally sit alongside other public-opinion and adoption-tracking work on generative AI, regulatory trust, and digital privacy attitudes once such notes are added.
Podcast
A research-radio episode discusses this paper: 🎧 MP3 · Spotify