Artificial intelligence (AI) has the potential to democratize expertise in medicine, bring expertise previously limited to specialists to a variety of health-care settings. But AI can easily falter, and making sure that AI works well across that variety of settings is a challenging task. Centralized governance, such as review by the Food and Drug Administration, can only do so much, since system performance will depend on the particular health-care setting and how the AI system is integrated into setting-specific clinical workflows. This Essay presents the need for distributed governance, where some oversight tasks are undertaken in localized settings. It points out the resource-based challenges of such governance and offers policy suggestions to ease the burden.
W. Nicholson Price,
Distributed Governance of Medical AI,
SMU Sci. & Tech. L. Rev.