Various scholars, commissions, and task-forces predict that Artificial Intelligence (AI) will transform and democratize health care systems, improving quality, safety, and efficiency, and offer the tantalizing possibility of empowering patients in customizing their health care experience. The use of AI has the potential to improve human health at the systems and service delivery levels. Well-designed algorithms could, for instance, correct presently high rates of avoidable medical errors. Yet, at the same time, the use of AI in health care presents a number of risks and implementation issues, such as issues of discrimination, informed consent, safety/quality (and liability for harm), and privacy. Failure to regulate in order to minimize these risks could undermine Canada’s ability to leverage the potential benefits of AI in healthcare. For instance, AI is unlikely to minimize medical error if it mirrors unrepresentative (and thus biased) training data that caused such errors. The use of poor quality AI in health care could, in turn, lead to public pushback and subsequent onerous regulations that stifle innovation. This “inconvenient truth” could result in lost opportunities for valuable AI developments likely to improve health care and the loss of financial and research efforts for unusable or underused AI.
With the support of the CIFAR AI + Society program, the “AI & Health Care: A Fusion of Law & Science” workshop team assembled an interdisciplinary group of experts in AI, law, ethics, policy, and medicine to address the core regulatory issue raised by these complexities: How can Canada maximize the potential benefits of the use of AI in health care while minimizing potential dangers? The group of invited experts began examining whether existing laws (constitutional protections, legislation/regulations, case law, and “soft” law (e.g. professional ethical and research ethics codes) are sufficient to protect patients from harms and maximize benefits that may emerge from the use of AI technologies and the possible reforms it may take.
The report AI & Health Care: A Fusion of Law & Science. An Introduction to the Issues presents the findings from a first event hosted in October 2020 which focused on the challenge of addressing privacy- and safety/quality-based concerns in legal regulations and related ethical issues surrounding the use of AI in healthcare. Highlighting a large number of potential regulatory issues, the report showcases the need for closer scrutiny of existing regulations and enquiry whether existing laws are well-suited to address the potential regulatory issues. The research team will begin analyzing these legal gaps in the next working session, which will focus on the regulation of AI in/as medical devices. That session will address safety and quality-related concerns but will also provide opportunities to further discuss privacy and discrimination.
The report is also available in French.
Led by Dr. Colleen M. Flood, the AI + Health lead and the University of Ottawa Research Chair in Health Law and Policy, the workshop was co-organized with Dr. Joelle Pineau, Associate Professor and William Dawson Scholar at McGill University and a core member of Mila, and Dr. Céline Castets-Renard, a CLTS Faculty member and the University of Ottawa Research Chair in Accountable AI in Global World.
The report was prepared by Dr. Michael Da Silva, the Alex Trebek Postdoctoral Fellow in AI and Health at the AI + Society Initiative. Dr. Florian Martin-Bariteau, University of Ottawa Research Chair in Technology and Society, and Dr. Teresa Scassa, Canada Research Chair in Information Law and Policy, also contributed to the report findings.
More details about our AI + Health research program can be found at aisociety.ca/health.