Knowledge is not all you need for comfort in use of AI in healthcare

The adoption of artificial intelligence (AI) in healthcare is rapidly expanding, transforming areas such as diagnostics, drug discovery, and patient monitoring. Despite these advances, public perceptions of AI in healthcare, particularly in Canada, remain underexplored. This study investigates the relationship between Canadians' knowledge, comfort, and trust in AI, focusing on key sociodemographic factors like age, gender, education, and income. Cross-sectional study. Using data from the 2021 Canadian Digital Health Survey of 12,052 respondents, we employed ordinal logistic and multivariate polynomial regression analyses to uncover trends and disparities. Findings reveal that women and older adults consistently report lower levels of knowledge and comfort with AI, with middle-aged women expressing the most significant discomfort. Comfort levels are closely tied to concerns over data privacy, especially regarding the use of identifiable personal health data. Healthcare professionals exhibited heightened discomfort with AI, indicating potential issues with trust in AI's reliability and ethical governance. Our results underscore that increasing knowledge alone does not necessarily lead to greater comfort with AI in healthcare. Addressing public concerns through robust data governance, transparency, and inclusive AI design is essential to fostering trust and successful integration of AI in healthcare systems.