American AI Sentiment in 2026
A · Strong evidence Gallup probability-panel survey of 3,152 US adults with a 2025 baseline and published margins of error, but a short descriptive report rather than analysis.
Summary based on the full text·Original: about 12 minutes to read
In short
The second annual survey by Gallup and the Special Competitive Studies Project (SCSP), published 15 September 2026, measures how Americans use, trust and judge AI. Use and self-reported knowledge rose over the year; trust did not, and economic optimism fell sharply, most among adults aged 18 to 29. On policy, respondents back education and workplace training and want safety rules and independent testing even if that slows development.
Key numbers
- 39% to 51%US adults using AI sometimes or more, 2025 to 2026
- 28%who fully or somewhat trust AI to make fair and unbiased decisions (32% in 2025)
- 45% vs 36%who expect AI to reduce versus boost US economic growth (53% boost, 24% reduce in 2025)
- 70%who expect AI to destroy more jobs than it creates over the next decade
As reported by the source.
English·PDF·opens scsp.ai
What it reports
- AI use "sometimes, often or always" rose 12 points, from 39% in 2025 to 51% in 2026; growth came mostly from people aged 30 and older (+16 points for 30-39, +13 for 40-64, +16 for 65+, +5 for 18-29).
- 68% call themselves very or somewhat knowledgeable about AI, up from 60%.
- Trust in AI to make fair and unbiased decisions: 28% (32% in 2025); 65% distrust; full trust stays at 2%.
- Trust among regular users is 43%, versus 12% among rare or non-users; trust among non-users fell from 23% to 12% in one year.
- 45% expect AI to reduce US growth, 36% to boost it, a reversal from 53% positive and 24% negative in 2025. Among 18-29 year olds the split is 57% reduce, 26% boost.
- 59% expect AI to destroy more businesses than it creates and 70% more jobs than it creates in the next decade; 20% expect most displaced workers to retrain successfully. Workplace productivity is the one positive: 52% expect a gain, 31% a loss.
- Learning is informal: 57% have asked an AI for help using AI versus 23% who asked a person; 26% used online videos or articles, 16% completed employer-provided training.
- 79% want safety and data security rules kept even if development slows (7% prefer speed); 75% want independent experts to do safety testing; 66% support on-the-job AI training programmes.
How they got it
Gallup Panel web survey of 3,152 US adults, 2-23 June 2026. The panel is probability-based (random-digit-dial and address-based recruitment), weighted to Current Population Survey figures. Margin of error is ±1.9 points around 50%, larger for subgroups.
Limits
- Use, knowledge and trust are self-reported; "AI use" is not defined by tool or task.
- US only; attitudes in Europe or Asia may differ.
- The report describes shifts but does not test why trust and optimism moved.
Why it matters for companies
The employees who will be asked to work with AI are using it more and trusting it less, and the youngest are the most pessimistic about jobs. Only 16% have had employer training while most learn from the tool itself, which leaves a gap companies can close directly. The trust gap between users (43%) and non-users (12%) suggests that structured, supervised use does more for acceptance than communication campaigns. Respondents' preference for transparency and independent testing is a useful guide for how internal AI rollouts are explained to staff.
Summary based on the full text. Read the original for the full argument and data.Open PDF on scsp.ai
- Tags
- Trust, AI adoption