Welcome to our SurveyMonkey AI Sentiment study, a quarterly report designed to measure ongoing changes in AI usage and consumer sentiment. This study began in Q4 2024, and updates will be released during the first month of each quarter. Below, you will find the latest results, historical results, and details on our research methodology.
Q3 2025 (July - September)
Our previous results are located here: Q2 '25 results | Q1 '25 results | Q4 '24 results
Canada and Australia see significantly higher AI adoption than the US and UK, but concerns linger
Nearly four in ten (39%) of adults in Canada and Australia have used AI within the past three months on a weekly or daily basis, compared with 30% of adults in Ireland, 35% in the UK, and 31% in the US.
Both countries also surpass the UK and the US in mixed feelings about AI’s impact on their lives, with four in ten adults in Australia (41%) and Canada (44%) who think AI will have both positive and negative effects on their lives, compared with 38% in the UK, Ireland, and US). Canada is also the only country where adults are significantly more concerned than impressed with the technology (35% vs. 29%).
AI usage in the US plateaus among frequent users, with 30% of adults using AI daily or weekly (on par with the previous quarter), but continues to make inroads among new users: 65% of adults say they have used AI at least a few times within the last three months, up from 61% in Q2, and 54% in Q4 of last year.
The leading AI use case across most countries is for work or professional tasks
Among adults who have used AI within the last 3 months: 51% of Australians use AI for work, on par with Canadians (47%) and significantly higher than adults in the UK (45%), and Americans (43%). Americans are significantly more likely than Australians, Canadians, and adults in the UK to use AI for fun or hobbies (39% vs. 32%, 34%, and 36%, respectively).
This month, in addition to our tracking survey, we also took a deep dive into AI usage and attitudes. Here’s what we found:
Though consumers may not actively be using AI-specific tools, most are using it passively
Almost all Americans are using some form of AI in their everyday lives, even if they are not aware. When asked about their usage of common tools that incorporate AI, 97% said that they use at least one of the following: streaming services (79%), navigation apps (78%), online shopping sites (78%), search engines (77%), social media (72%), voice assistants (47%), predictive text (42%), or smart home devices (39%). Even among those who use them, not everyone understands that these tools incorporate AI. Most (86%) search engine users understand that search engines incorporate AI, along with 81% of voice assistant users. However, only 75% of social media users understand that social media incorporates AI, in line with 74% of predictive text users, and higher than the 66% of navigation app users, 65% of online shoppers, 64% of smart home device users, and 59% of streaming service users. The pattern is similar across the UK & Ireland, Australia, and Canada.
Comfort levels with AI vary by use case
Consumers are more comfortable with AI being used in lower-stakes situations. Half (55%) of American consumers are comfortable with an AI agent taking their order at a drive-thru, while only a third are comfortable with AI being used by HR to scan through resumes (34%), investment accounts using AI to make choices about trades (32%), or their doctor using AI to personalize medical advice (33%), and even fewer are comfortable riding in a self-driving car (25%). Those who are uncomfortable tend to be ‘very uncomfortable’ while those who are comfortable tend to only be ‘somewhat comfortable’; for example, 40% are very uncomfortable with their doctor using AI, while only 9% are very comfortable, and 51% are very uncomfortable riding in a self-driving car, while only 9% are very comfortable. Among workers who can use AI at work, 65% are comfortable using AI in their day-to-day work. Comfort levels with AI use cases are similar across countries.
Employees want AI training, but few are offered it
Only 13% of American workers say that their company offers them any AI training (63% say they are not offered training, and 23% are not sure). However, most workers are interested in building AI skills. While only 13% have already participated in AI training, another 52% would participate if it were offered by their employer. Older workers are just as invested as their younger colleagues, if not more; 62% of Gen Z workers, 61% of Millennials, 68% of Gen X, and 72% of Boomers/Silent Gen have either taken AI training, or would take it if offered.
Americans are significantly less interested in AI training than other countries – 35% of US workers say they wouldn’t be interested if training was offered, vs. 28% of Australians, 29% of Canadians, and 27% of those in the UK and Ireland.
Methodology: This quarter’s SurveyMonkey study was conducted July 11, 2025 to September 23, 2025 among a sample of 10,558 adults in the US, 1,318 adults in Canada, 171 adults in the Republic of Ireland, 2,024 adults in the UK, and 1,235 adults in Australia. Respondents for this survey were selected from a non-probability online panel. The modeled error estimate for each country is shown in the table below. U.S data have been weighted for age, race, sex, education, and geography using the Census Bureau’s American Community Survey to reflect the overall demographic composition of the United States. Non-US data have been weighted for age and sex based on each country’s latest census data.
| Period | Sample size | Fielding dates | Margin of Error (MOE) |
| Q3 ‘25 | US - n=10,558 adults CA - n=1,318 IE - n=171 UK - n=2,024 AU - n=1,235 | 07/11/25 - 09/23/25 | US - +/- 1.0pp CA: +/- 3.0pp IE: +/- 10.0pp UK: +/- 2.5pp AU: +/- 3.0pp |
| Q2 ‘25 | n=6,997 U.S. adults | 04/11/25 - 06/23/25 | +/- 1.5pp |
| Q1 ‘25 | n=41,373 U.S. adults | 01/22/25 - 03/31/25 | +/- 1.0pp |
| Q4 ‘24 | n=25,030 U.S. adults | 10/25/24 - 12/31/24 | +/- 1.0pp |
The supplementary survey was conducted from September 23 - October 2, 2025 among a sample of 1,037 adults in the US, 931 in the UK and Ireland, 551 in Canada, and 570 in Australia. Respondents for this survey were selected from a non-probability online panel. The modeled error estimate for this survey is plus or minus 2.0 percentage points. US data have been weighted for age, race, sex, education, and geography using the Census Bureau’s American Community Survey to reflect the overall demographic composition of the United States. International data have been weighted for age and gender using data from the Organization for Economic Co-operation and Development (OECD).
