The use of generative AI and large language models is currently an immature field. Although the possibilities are many, there is also great uncertainty about risks related to how this type of AI can be used in a trustworthy way in health and care services. We simply lack insight into the side effects.
However, the field is developing rapidly, with among other things mechanisms to improve, ground in knowledge, and gain better control over the quality of AI systems with generative language models.
One central measure to reduce risks and increase quality is to develop and make available language models, i.e., foundation models, that are adapted to conditions in the Norwegian health and care sector, and that industry and the public sector can build on for their use. This is in line with the national digitalization strategy which states that the Government will work to further develop a national infrastructure for AI that will provide access to foundation models based on Norwegian and Sami languages and social conditions [95].
The next chapters cover how large language models can be adapted to Norwegian health and care services (Chapter 4) and propose measures that together will facilitate the use of generative AI models being based on foundation models adapted to conditions in the Norwegian health and care services (Chapter 5).
[95] https://www.nb.no/pressemeldinger/regjeringen-vil-satse-stort-pa-digitalisering-og-kunstig-intelligens-ved-nasjonalbiblioteke In the State Budget for 2025, 20 million kroner is allocated to increased investment in the National Library's work with artificial intelligence and training of generative language models.