Artificial Intelligence in Healthcare: Can Regulation Catch Up with Innovation?

Effective regulation of AI remains challenging due to several unique aspects of these technologies and their applications to healthcare. First, AI innovation is affecting a number of diverse segments of healthcare that face different regulatory risks. Innovators are looking to lend AI to areas such as clinical decision support, utilization review, reimbursement and payment, and research, among many others. This makes it difficult to establish a one-size-fits-all regulatory framework. Second, even within a specific segment of the industry, AI solutions are being created for an array of different purposes, which further frustrates creating a unified approach to regulation. Third, even if the absence of this variability, the definition of AI remains debated among experts and regulators. Without a common understanding about the attributes of these technologies that create risk, stakeholders will have difficulty determining what aspects of AI warrant regulation. Finally, many AI technologies are developed and operate as “black boxes” with opaque processes, often with the capacity to engage in unforeseeable actions. Without a robust understanding of how these technologies function, regulators will have difficulty developing guardrails for responsible development and use of AI.