Community Health Nurses' Knowledge and Perceptions of AI in Canada: National Cross-Sectional Survey.

Background: Artificial intelligence (AI) continues to expand into nursing and health care. Many examples of AI applications driven by machine or deep learning are in use. Examples include wearable devices or alerts for risk prediction. AI tends to be promoted by nonnurses, creating a risk that AI is not designed to best serve registered nurses. Community health nurses (CHNs) are a small but essential group. CHNs' familiarity with AI and their perceptions about its effect on their practice are unknown. Objective: The research aims to understand CHNs' awareness, knowledge, and perceptions of AI in practice and gain insights to better involve them in AI. Methods: An online cross-sectional Canadian survey targeting CHNs was conducted from April to July 2023. Descriptive statistics summarized respondents' characteristics and perceptions of AI, followed by a chi-square test used to determine a relationship between respondents' level of AI knowledge and their AI perceptions, with odds ratio (OR) to determine the strength of association. Results: A total of 228 CHNs participated with varying response rates per question. Most respondents were female (172/188, 91.5%), average age of 45.5 (SD 11.7) years, and an average of 13.5 (SD 10.1) years of community practice experience. Most respondents (205/228, 89.9%) felt they welcomed technology into their practice. They reported their understanding of AI technologies as "good" (95/220, 43.2%) and "not good" (125/220, 56.8%). Overall, 39.6% (80/202) of respondents felt uncomfortable with the development of AI. They agreed that AI should be part of education (143/203, 70.4%), professional development (152/202, 75.2%), and that they should be consulted (195/203, 96.1%). Many respondents had concerns related to professional accountability if they accepted a wrong AI recommendation (157/202, 77.7%) or if they dismissed a correct AI recommendation (149/202, 73.8%). Respondents with "good" AI knowledge were significantly associated with, and twice as likely to indicate nursing will be revolutionized (P=.007; OR 2.28, 95% CI 1.25-4.18), nursing will be more exciting (P=.001; OR 2.52, 95% CI 1.42-4.47), health care will be more exciting (P=.004; OR 2.3, 95% CI 1.30-4.06), and agreed that AI is part of nursing (P=.01; OR 2.1, 95% CI 1.19-3.68). Respondents with "not good" AI knowledge were significantly associated with, and more likely to feel uncomfortable with AI developments (χ21=4.2, P=.04; OR 1.84, 95% CI 1.03-3.3). Conclusions: CHNs reporting "good" AI knowledge had more favorable perceptions toward AI. Overall, CHNs had professional concerns about accepting or dismissing AI recommendations. Potential solutions include educational resources to ensure that CHNs have a sound basis for AI in their practice, which would promote their comfort with AI. Further research should explore how CHNs could be better involved in all aspects of AI introduced into their practice.