Elvin R. Latty Professor of Law Arti Rai, who co-directs the Duke Center for Innovation Policy, and colleagues at the Duke University’s Robert J. Margolis, MD, Center for Health Policy (Duke-Margolis Center) are studying an emerging problem in artificial intelligence (AI)-enabled health care delivery: the tension between the need for “explainability” of treatment rationales versus the need to protect trade secrets in the burgeoning area of clinical decision support software innovation (CDSS).
Health care professionals’ duty to promote patient welfare includes basing decisions on sound and explainable rationales. Institutional providers, federal regulators, and payors also need information relevant to explainability. But certain types of detailed explanations could potentially facilitate reproduction and in that way, compromise trade secrets, a key incentive for innovation in the field.
Rai, an internationally recognized expert in intellectual property law, innovation policy, administrative law, and health law, said the project is the first of its kind to tie explainability issues to quantitative data on current needs for trade secrecy by commercial actors.
“While software development has always involved trade secrecy, the importance of trade secrecy as an innovation incentive may have increased as a consequence of challenges associated with securing and enforcing software patents,” she said. “For this reason, the principal regulator of AI-based software in health care, the FDA, as well as professional organizations, providers, and insurers are actively interested in the question of how to balance explainability and trade secrecy.”
Rai is a principal investigator on the project, which is funded by a $196,000 “Making a Difference” grant from The Greenwall Foundation. Rai had previously co-authored a white paper with Gregory Daniel, PhD, MPH, deputy director, policy and clinical professor at the Duke-Margolis Center and Christina Silcox, PhD, a managing associate at the center. That prior paper serves as a background document on the legal and regulatory landscape surrounding AI-based CDSS.
The current project, which will gather information through private workshops, public conferences, and databases of patents and venture capital funding, will produce peer-reviewed journal articles as well as white papers.