Addressing the Trust Gap
The Role of Virtual Services and AI in Mental Health Care Across Canada
National polling shows AI-enabled mental health tools are already being used, and many users are finding them effective.
AI-enabled tools and virtual services are already part of how people in Canada are looking for mental health support, but confidence has not caught up.
While public trust in AI-enabled tools remains low, the experience of users tells an important story. Many people who have used AI-enabled tools or virtual mental health services report that they find them effective. The opportunity now is to strengthen trust through evidence, informed consent, accountability and a continued focus on human-centred care, while recognizing the role technology can play in helping people access support.
The Role of Virtual Services and AI in Mental Health Care Across Canada examines how people living in Canada are using and perceiving digitally-supported mental health tools, including AI-enabled tools and virtual mental health services.
The report, prepared in partnership with the Mental Health Commission of Canada, points to a clear opportunity: AI-enabled tools and virtual services may help expand access to support, but public confidence depends on evidence, transparency and a human-centred approach.
Key Findings
Approximately six million people in Canada have used AI-enabled tools for mental health support in the past year: AI-enabled tools are being used for general well-being, companionship and to help manage mild-to-moderate stress, showing that many people are already turning to them for lower-intensity support.
Many users report positive experiences: Three out of four people who used AI-enabled tools or virtual mental health services found them effective in supporting their mental health or well-being.
Public trust remains low: Only 14% of people in Canada say they trust AI tools to provide helpful and safe mental health support.
Virtual care is already part of the mental health landscape: Among people who accessed mental health care in the past year, 45% did so virtually, either entirely virtually or in combination with in-person care.
AI-enabled tools may be a bridge to care for some users: Forty percent of AI users said using AI made them more likely to seek professional mental health support.
Why this matters
Digitally-supported mental health tools may help address some barriers to care, including cost, availability, privacy and access outside traditional service hours. They may also help some people take a first step toward support.
But technology should support human-centred care, not replace it.
As AI-enabled tools and virtual services become more common, ongoing research is needed to understand where they help, where risks remain and what safeguards are required. The findings reinforce the importance of tools that are safe, effective and evidence-based, while ensuring that people, relationships and professional care remain at the centre of the mental health care experience.
ABOUT THE REPORT
This report draws on MHRC’s national population polling initiative and provides timely insight into how people living in Canada engage with, understand and perceive virtual services and AI-enabled tools in mental health care.
The findings are intended to support evidence-informed discussion about the role of digitally-supported tools in the mental health system, including how they can be used safely and equitably alongside human-centred care.
METHODOLOGY
The findings are based on an online survey of 3,519 adults living in Canada, conducted in February 2026 with the support of Pollara Strategic Insights. This was the twenty-seventh poll in MHRC’s national population polling series, which was launched in April 2020 to monitor the mental health of people living in Canada.
The study was conducted as a blind survey, meaning participants were unaware of the subject matter before beginning the survey. National results were weighted using the most recent census data on gender, age and region to ensure the total sample was representative of the population.
Results were rounded to the nearest whole number, and findings were not reported for sample sizes below 10. The report does not include statistical significance test results.