Wissenschaftliche Begleitung für die Beobachtung der Behördenkooperation im Rahmen des Pilotprojekts Reallabore für die KI-VO

The project examines how AI regulatory sandboxes established under the EU AI Act can be designed as cooperation and learning arrangements among authorities with decentralised responsibilities. It focuses on the scientific observation of a pilot project involving the future notifying authority, market surveillance authorities, Accreditation Austria and RTR GmbH. The aim is to identify cooperation needs, potential conflicts of competence and organisational interfaces and to derive evidence-based contributions for the future design of regulatory sandboxes. The project combines legal and organisational analysis, stakeholder mapping, comparative European and international case studies, structured interviews, accompanying workshops and qualitative content analysis. It examines divergent legal interpretations, communication and knowledge gaps, coordination burdens and enabling factors for effective inter-authority cooperation. Findings will be presented in an interim report, validated with the participating organisations and consolidated in a final report. The project will provide structured contributions to recommendations concerning legally robust, efficient and operationally viable cooperation, communication and knowledge-building processes, financing arrangements and the possible involvement of additional actors from Austria’s AI ecosystem.