AI COMPETITIONS AS INFRASTRUCTURES: EXAMINING POWER RELATIONS ON KAGGLE AND GRAND CHALLENGE IN AI-DRIVEN MEDICAL IMAGING

Abstract

Artificial Intelligence (AI) is quickly being taken-up across scientific disciplines, medical imaging is no exception. To stimulate development and facilitate the scientific evaluation of new approaches, AI-based research in medical imaging is increasingly organised in a competitive manner through digital machine-learning development platforms such as Kaggle and Grand Challenge—two of the leading platforms in the field. For medical image analysis, such competitions constitute an important research infrastructure, steering global research and development in this dedicated AI subfield. Yet, little is known about how these platform-based infrastructures that operate across the medical AI research pipeline shape the conditions for model production and evaluation. This paper addresses this issue through a critical empirical case study of 120 medical imaging competitions on Kaggle and Grand Challenge between 2017 and 2022. We show that platforms as well as competition organisers shape power relations in medical AI research at the level of data and task design, model production and evaluation in several distinct ways. Taken together, because competitions play a central rol within the field, these findings highlight the impact these powerful actors ultimately have on steering medical image AI research directions as they see fit and influence the types of models that are implemented into clinical settings.