Accuracy in healthcare AI and its associated harms. Reframing existing concepts to renew forms of patients’ redress

Abstract

The progressive integration of AI-driven systems into medical devices has enabled new ways of functioning, enhanced by AI methods and big data. In this context, medical device failures assume new nuances. Not only technical failures, but also manifestations of their inherent epistemic limits, biases, and logics of AI-based knowledge production are now in the harm equation. In this complex landscape, patients face new risks, including from (in)accuracy of data and AI. As part of the issue ‘The Failed Epistemologies of AI? Making Sense of AI Errors, Failures and their Impacts on Society’, this paper argues that the EU legal framework necessitates a reconceptualisation of the notion of accuracy in AI-based healthcare technologies, to equip patients with broadened means of redress. Therefore, the article (i) proposes new ways of considering the concept of accuracy in EU law; and (ii) comments on harm and the viable redress options for patients. In light of the analysis, the article notes that the digitalisation shift in healthcare implies the emergence of new redress mechanisms and that existing legal frameworks should be revised or supplemented with new legal tools to address healthcare AI-associated inaccuracies.

Ähnliche Arbeiten