AI Needs High-Quality Health Data at Scale - Will the EHDS Deliver?

Abstract

A panel discussion was held during the DigiHealthDayS-2024 Scientific Congress on 15^th November 2025. It explored the potential for the European Health Data Space (EHDS) to become a high-value resource of good quality data for the development of AI innovations to support better and safer healthcare. Panel members discussed whether the measures presently intended for the provision of secondary used data sets within the EHDS and AI Act will be suitable for AI development. At present, a standardized data quality label will be defined, but its use will be optional, interoperability standards for data sets sharing for research are not mandated. However, AI innovators need access to large-scale, high-fidelity data sets that have well-documented provenance and quality, and which are accurately representative of the populations on whom the innovators wish to target their solutions. The EHDS has the potential to accelerate the availability of high-quality data sets, but the adoption of the data quality and utility label to assess the data sets must be strongly promoted and accompanied by measures and incentives for health systems to actively improve the quality of the data they routinely collect within EHR systems.