Responsible Adoption of Artificial Intelligence (AI) in Pharmacy Practice: Perspectives of Regulators in Canada and the United States

BACKGROUND: Use of Artificial Intelligence (AI) is proliferating in society and in pharmacy practice. For some, this represents a great advancement that will enhance effectiveness and efficiency of health care. For others, it is an existential risk that will worsen inequalities, lead to deskilling of the workforce, and spiral beyond the comprehension or control of humans. Human-in-the-loop (HiL) vs. human-out-of-the loop (HoL) AI have different potential risks and challenges that raise questions regarding patient safety. Defining principles for responsible adoption of AI in pharmacy practice will be an important safeguard for both patients and the profession. METHODS: Semi-structured interviews with 12 pharmacy regulators from across Canada and the United States were undertaken, with informed consent. Constant comparative data analysis using nVivo v15 was used to identify common themes. The COREQ framework was applied to assure quality of research processes used. RESULTS: Pharmacy regulators highlighted the value of a principles-based, rather than rules-based, approach to AI. Core principles related to transparency, redundancy, audit and feedback, quality assurance, privacy/data security, alignment with codes of ethics, and interoperability were identified. There was limited consensus on the role of consent and choice as principles to be considered. CONCLUSIONS: The role of regulation in shaping responsible adoption of AI in pharmacy will be significant. This study highlighted a series of agreed-upon principles but also identified lack of consensus with respect to how consent and choice could be operationalized in pharmacy practice.