From user to manufacturer: How substantial modification of medical AI crosses the product liability threshold under the revised PLD

The revised Product Liability Directive (rPLD) introduces a pivotal shift in the liability landscape for healthcare professionals using artificial intelligence. While existing research often focuses on professional or regulatory thresholds, this paper highlights a critical and underexplored juncture: the moment a health institution or professional's interaction with an AI system constitutes a substantial modification, thereby transforming them from a user into a de facto manufacturer under the rPLD's strict liability regime.
Drawing on the author's prior work on threshold-based liability frameworks, this paper examines how the rPLD redefines the “product” to include software and AI systems, and expands liability to any natural or legal person who substantially modifies a product outside the manufacturer's control. In clinical practice, such modifications may occur through actions such as overriding safety parameters, integrating unauthorized components, or in other manner. When these actions are not foreseen in the manufacturer's initial risk assessment, create a new hazard or increase the risk level, and are performed by a healthcare professional rather than a consumer, the health institution or professional crosses the product liability threshold, triggering strict liability for any resulting harm.
The analysis situates this threshold within the broader EU regulatory ecosystem, including the AI Act, the General Product Safety Regulation, and the Medical Devices Regulation, to demonstrate how liability becomes distributed and context-dependent. The paper argues that the rPLD does not merely complement existing fault-based regimes but creates a distinct, strict liability pathway that reallocates risk to the party with the highest degree of control over the AI system's final safety configuration.