Titolo: Beyond compliance: AI Act made usable in healthcare
ACRONIMO: USEAI
Responsabile scientifico: prof. Claudio Sarra - Dipartimento di Diritto Privato e Critica del Diritto – Università degli Studi di Padova
Team di Progetto:
Prof. Claudio Sarra, Dipartimento di Diritto Privato e Critica del Diritto – Università degli Studi di Padova
Prof. Nicolò Navarin, Dipartimento di Matematica “Tullio Levi-Civita” – Università degli Studi di Padova
Prof. Anna Spagnolli, Dipartimento di Psicologia Generale – Università degli Studi di Padova
Bando: BAC Partenariato Esteso FAIR “Future Artificial Intelligence Research”, obiettivo 9 AI Act Analysis, bias and applications in healthcare – SPOKE 8 Alma Mater Studiorum Università di Bologna
Durata: 01/04/2024 - 31/03/2025 (12 mesi)
Budget totale progetto: € 102.377,00

Abstract del progetto
The project aims to provide actionable guidelines for legally and ethically compliant data processing in healthcare for administrative, clinical, surgical, and insurance purposes. The objective is to disclose and mitigate the risk of potential biases and discrimination and promote trustworthiness in healthcare AI. These objectives are coherent with Objective 9 of Spoke 8. The approach is an innovative, interdisciplinary methodology combining Computer Science, Law, and Psychology where the competencies of the team members mutually inform each other and converge on the production of compliance information guidelines targeting the specific applications of AI in healthcare. Three synergic activities will be started in month 1: a scoping review of data and algorithm fairness strategies, the analysis of relevant EU regulations on AI, and an empirical study of the stakeholders' information needs regarding AI compliance. Subsequently, a checklist of actional legal requirements for AI-based data processing will be outlined, and, on its basis, the legal adequacy of the data and algorithm fairness strategies will be assessed, and a second study with stakeholders will be run. In the last two months of the project, the results will be wrapped up, and a set of resources will be created, including a checklist of actionable legal requirements for compliance and related explanatory material. The foreseen impact is the removal of barriers to AI Act adoption through simplification of legislation and advancement in the methods to achieve transparency in the use of AI to the AI stakeholders in healthcare, including external evaluators.