AI use is surging in Canada, with nearly half of Canadians (47%) reporting they’ve used AI tools—up from just 25% in February 2023. This growth reflects AI’s increasing integration into both professional and personal settings. However, while usage climbs, trust and regulation remain key issues. Leger has tracked these questions for the last three years.
Key findings
- Nearly half of Canadians have tried AI tools.
Among Canadians, 23% have used AI for work or school and 36% in a personal context. Young Canadians (18–34) lead the way, with 73% having used AI tools. - Canadians are largely satisfied with their AI experiences.
Among users, 72% rate their experience as either “good” or “excellent”. Satisfaction is highest among younger adults and urban residents. - Trust in AI varies widely depending on the task.
Canadians are most likely to trust AI with:
- Home tasks like adjusting thermostats (53%)
- Answering product questions via chat (52%)
- Using facial recognition for access (46%)
However, trust plummets for:
- Driverless transport (20%)
- Teaching children (14%)
- Helping find a life partner (14%).
- Opinions are divided on AI’s societal impact.
Only 32% of Canadians believe AI is good for society, a point which has been trending downwards over the last three years. 35% think it’s bad, and another third remain unsure. - Privacy and dependency top the list of concerns.
Canadians’ biggest concerns about AI include:
- Overdependence on tools (40% top 2 concerns)
- Lack of emotion/empathy (28%)
- Threats to jobs (35%).
- Regulation is a near-unanimous priority.
An overwhelming 86% of Canadians support government regulation to ensure AI is used ethically and safely, with 56% strongly in favour. - Impact on work is expected—but not always disruptive.
Among working Canadians, 82% expect AI to impact their industry in the next year, though only 31% foresee a “significant” impact.
Methodology
The results for this wave of research are based on online research conducted from March 30 to March 30, 2025, with a representative sample of 1,628 Canadian adults 18 years of age and older from Leger’s LEO panel. The data was statistically weighted according to 2021 Canadian Census figures.
A margin of error cannot be associated with a non-probability sample in a panel survey, but for comparison purposes, a probability sample of 1,628 would have a margin of error of +/- 2.43%, 19 times out of 20.
