Surveys in Canada and the United States show more caution than enthusiasm about AI. In a Pew Research Center survey fielded June 9 to 15, 2025, half of U.S. adults said AI in daily life made them more concerned than excited and 10 percent said more excited. Ipsos data reported in Stanford's 2025 AI Index put the share saying AI products offer more benefits than drawbacks at 40 percent in Canada and 39 percent in the United States in 2024, against 55 percent worldwide.
Key takeaways
- Pew found 50 percent of U.S. adults more concerned than excited about AI, 38 percent equally both, and 10 percent more excited.
- Support is high for narrow uses such as weather forecasting (74 percent) and low for personal ones.
- Canada rose from 32 percent to 40 percent seeing more benefit than drawback between 2022 and 2024, and the United States from 35 percent to 39 percent.
- Opinion measures attitudes, not outcomes, and surveys with different wording cannot be compared directly.
What the Pew survey measured
The Pew Research Center report, published September 17, 2025, draws on 5,023 U.S. adults in the American Trends Panel. Half said the growing use of AI in daily life made them more concerned than excited, up from 37 percent in 2021. Thirty-eight percent were equally concerned and excited, and 10 percent were more excited than concerned. Fifty-seven percent rated the risks of AI for society as high, while 25 percent rated the benefits as high. About six in 10 said they want more control over how AI is used in their own lives, and nearly three-quarters said they would let AI help with at least some day-to-day tasks.
The same respondents were not uniformly negative, which is what makes the data useful. Concern about society sits next to willingness to use specific tools.
Where majorities see a role for AI
Pew asked whether AI should play at least a small role in particular jobs. Seventy-four percent said yes for forecasting the weather, 70 percent for searching for financial crimes, 70 percent for searching for fraud in government benefits claims, and 66 percent for developing new medicines. The pattern is that people accept AI where it analyzes large volumes of data for a defined purpose and are more reluctant where judgment about personal life is involved.
That preference lines up with the kinds of measured gains documented elsewhere, such as faster drafting in controlled experiments, but opinion about a use is not evidence that the use works.
Where people expect harm
Respondents were asked how AI will affect human abilities. Fifty-three percent said it will worsen people's ability to think creatively, compared with 16 percent who said it will improve it. On forming meaningful relationships, 50 percent expected a worse outcome and 5 percent a better one. Problem-solving was closer: 38 percent said worse and 29 percent said better. These are expectations, not measurements of skill, and respondents could be wrong in either direction.
Canada, the United States, and the world
The Ipsos series in the 2025 AI Index public opinion chapter asks whether AI products and services have more benefits than drawbacks. Pew's global survey of 25 countries, published October 15, 2025, used a different question and reported a median of 34 percent more concerned than excited, 42 percent equally both, and 16 percent more excited. Its Canadian figure was not available on the page reviewed, so the Ipsos series is the Canadian reference here.
| Place | More benefits than drawbacks, 2022 | 2024 | Change |
|---|---|---|---|
| Canada | 32 percent | 40 percent | Up eight points |
| United States | 35 percent | 39 percent | Up four points |
| Global | 52 percent | 55 percent | Up three points |
Trust in institutions also differs. Pew found that 44 percent of U.S. adults trusted their country to regulate AI effectively and 47 percent did not.
How to read an AI opinion poll
- Check the exact question: "concerned versus excited" and "benefits versus drawbacks" measure different things.
- Look at the field dates, because the products people picture change quickly.
- Separate attitudes toward AI in general from attitudes toward a named use.
- Note who was sampled, since a U.S. panel says nothing about Canadian adults.
- Treat a poll as a record of public mood, not as proof that AI is helpful or harmful.
FAQ
Do most Americans see AI as a benefit to humanity?
Not in the Pew data. Ten percent were more excited than concerned, and 25 percent rated the benefits of AI as high, though many accept specific uses.
Are Canadians more optimistic than Americans?
Slightly, by one measure. In the 2024 Ipsos figures, 40 percent in Canada and 39 percent in the United States said AI has more benefits than drawbacks, a gap small enough that it may not be meaningful.
Has opinion improved over time?
On the Ipsos measure it has risen in both countries since 2022. On Pew's concern measure it moved the other way in the United States, from 37 percent to 50 percent more concerned between 2021 and 2025.
Does public opinion decide whether AI is good for society?
No. Opinion shows what people expect and want, while outcomes need studies with comparison groups.
Related reading
- Threat or tool: the continuing debate
- Arguments against the threat view
- Whether AI can dream
- Counting the harms: incident data on AI
More in AI Policy and Ethics.
Last verified: October 2026
Sources
- How Americans View AI and Its Impact on People and Society, Pew Research Center, September 17, 2025
- How People Around the World View AI, Pew Research Center, October 15, 2025
- Public Opinion chapter, AI Index Report 2025, Stanford Institute for Human-Centered AI, 2025 (Ipsos data)