Axxessment:AI

Offentlig prosjektsammendrag

Big data and artificial intelligence are having a noticeable impact on most aspects of life, from work to transport to commerce. The success of these various applications rests heavily on the quality and quantity of data available to develop models, the robustness and performance of these algorithms, and the trustworthiness of the systems for the stakeholders affected. In the health system, while such tools have enormous potential in improving the efficiency and quality of care, their adoption in routine clinical settings remains limited. While the field faces technical challenges, regulatory uncertainty is commonly cited as a barrier preventing their uptake. Software developers are unclear about the regulatory process, regulatory bodies are uncertain about how existing regulations apply to challenging aspects of these new technologies, and clinicians and hospital staff are left wondering if a given AI system can be safely put into service. This project explores the safety and quality of AI-based medical technologies, and how these topics relate to emerging regulatory requirements. It will develop an assurance framework and supporting IT platform to aid in the assessment of AI-based medical technologies. This framework will include a structured knowledge-base developed through stakeholder interviews, workshops, and in-depth analysis of regulatory documents and scientific literature. In parallel, a software prototype will be developed and tested with the goals of aiding assessors as they evaluate medical technologies. Taken together, this prototype should increase clarity and provide a framework for the assessment of AI-based medical technologies, resulting in a more efficient and consistent regulatory process, and ultimately greater safety and trust.