AI in Medicine: from Expert Systems to Deep Learning and Algorithm-driven Health Care

• In Event: The Prediction Factor: Medical Decision in the Age of Big Data - II: Concerns about the prediction factor: regulation in question?

Abstract

Abstract

The present paper aims at highlighting the historical processes underpinning contemporary excitement about artificial intelligence (AI) and algorithm/data-driven innovation in healthcare. We focus on the tumultuous trajectory of AI’s hype and disillusionment and its social construction (as well as negotiations about its validity as a pure pattern); the lessons or potential pitfalls from the international arenas of second-generation AI (roughly between the mid-1980s and late 1990s) which may benefit contemporary adopters and vendors in and around medical practice. How much can we learn from the disillusionment following the enthusiasm with knowledge-based expert systems? What did we learn from DARPA’s Strategic Computing Initiative? What was the legacy of the European Union’s ESPRIT programme? What examples are there to date that show AI and algorithm/data-driven innovation in healthcare has still a long way to go to be ethically sound and produced to gain health equality/equity? We draw from historical and contemporary literature as well as our empirical research to shed light on issues of classification, accuracy, and in/equality in currently proposed AI technologies for precision medicine. By doing so, we aim at offering initial glances towards a responsible framework, informed by STS approaches, currently missing from existing debates around responsibility/accountability.

Authors

• N. Elida Detfurth, York University, Canada

• Renate Baumgartner, Eberhard Karls Universität Tübingen

• Vassilis Galanos, University of Edinburgh