Clicking Away Consent: Establishing Accountability and Liability Apportionment in Direct-to-Consumer Healthcare Artificial Intelligence

Authors

Stephanie L. Lee

Abstract

Artificial Intelligence (AI) and machine learning are making sweeping changes across all industries, and health care is no exception. AI promises to revolutionize patient treatment with the development of algorithm-driven tools to improve efficiency in clinical care. As alluring as machine-driven learning may be given its potentialities, however, the incorporation of AI into the healthcare field has also been received with trepidation. This fear is understandable given the lack of transparency to the public surrounding the exact mechanisms for creating algorithms and the reasoning followed by the software. Indeed, AI in the healthcare system is aptly known as “black-box medicine.” Liability apportionment for when AI malfunctions or errs is special cause for concern and an area of tort law that remains largely uncharted, controversial, and jurisdiction-dependent. This note sets out to provide solutions to the moral and legal concerns raised by AI developers’ use of liability waivers to escape culpability, specifically in the context of direct-to consumer health and medical mobile applications (apps). The seminal case of Tunkl v. Regents of University of California provides an excellent six-factor framework for the judicial interpretation necessary to ensure accountability and transparency in AI-based treatment. This note argues that the Tunkl factor of bargaining power dynamics should hold controlling weight as it deals with the essence of contract validity based on mutual understanding and voluntariness. Focusing on the power dynamic between AI developers and consumers is especially important in the healthcare app space because of the heavy influence of information asymmetry and heuristic biases, which are exacerbated by the black box nature of medicine. This framework should be applied in invalidating exculpatory clauses contrary to public policy and in finding all waivers of liability required to be signed by app users to be presumptively invalid, regardless of the categorization of the app (i.e., for both health apps and wellness apps). Centering concerns over bargaining power and voluntariness in judicial interpretation of exculpatory clauses, and thus upholding the principles of contractual freedom, is essential to establishing accountability and transparency in AI healthcare use.

Recommended Citation

Stephanie L. Lee, Clicking Away Consent: Establishing Accountability and Liability Apportionment in Direct-to-Consumer Healthcare Artificial Intelligence, 88 Brook. L. Rev. 1355 (2023).
Available at: https://brooklynworks.brooklaw.edu/blr/vol88/iss4/7