Responsible use of AI

Important results from the project

the project objectives were achieved. It developed seven principles for responsible AI and four solution proposals, grounded in target groups’ needs and positively received. Internal AI expertise was also strengthened. A key insight was that AI’s value depends on the entire system around the technology, not the technology alone. The project also revealed system gaps without clear ownership, including inadequate follow-up and unclear patient information about AI.

Expected long term effects

In the longer term, the project is expected to increase trust in AI in healthcare through clearer governance, accountability and monitoring. This may reduce regional disparities, strengthen patient information and improve patient safety. Greater national coordination between authorities could enable faster and safer AI adoption. The results also provide a foundation for continued work on Trust in AI in Health and Social Care, building on the study’s findings.

Approach and implementation

The project was carried out by the National Board of Health and Welfare, AI Sweden and Region Örebro County, with funding from Vinnova. The work used service design and systems thinking, with breast cancer screening as a pilot case. It included a literature review of 159 sources, in-depth interviews with 15 people across 8 regions, a design sprint, systems analysis and a future-oriented seminar with researchers. The work resulted in seven principles and four tested solution proposals