Does artificial intelligence disrupt or does it accelerate? Path dependency in medical innovation

• In Event: 30416 - The Culture and Politics of Emerging Technology

Abstract

Discourse on AI in medicine often takes a deterministic approach, focusing on whether its transformative effect will be positive or negative, rather than examining the underlying forces shaping its development. This study empirically analyzes AI innovation in medical technology from 2010-2023 using data from patents and the Food & Drug Administration. We show AI innovation in healthcare is constrained by larger existing social and economic trends. Using time-series models and vector embeddings, we provide visualizations of the nature of AI in medical technology development. Contrary to narratives that frame AI as a general-purpose technology destined to transform the whole of the medical system, we show that development is concentrated in a few key areas, particularly radiology, and develop a model on what motivates individual firms to produce AI-based technology. Rather than fundamentally transforming the medical technology industry, AI innovations have so far primarily accelerated existing trends, and building on the imaging turn in medicine by reinforcing the growth of well-established fields such as radiology. In this way, we argue that the growth of AI in medical technology is not purely driven by technological superiority or economic rationality, but that it reflects the path-dependent nature of innovation.

Author

• Anne Kavalerchik, Indiana University-Bloomington