Goals

1. Map the legal, regulatory, and contractual frameworks governing hospital data use and
   commercialization of AI.
2. Assess the impact of the lärarundantag on hospital access to AI innovations.
3. Evaluate existing contractual mechanisms and conduct risk-based analysis.
4. Benchmark international practices to identify effective models.
5. Explore integration of open science principles with fair benefit-sharing.
6. Develop practical collaboration models (shared ownership, preferential access, revenue-sharing).
7. Formulate policy recommendations for sustainable data-sharing partnerships.

Research questions

• How do existing Swedish and EU legal frameworks govern the use of hospital health data in AI development?
• What specific legal tensions arise from the interaction between the lärarundantag and hospital interests in securing access to AI innovations?
• How do open science policies interact with IP rights and commercialization pathways in this context?
• How are data-sharing and collaboration agreements currently structured between hospitals, universities, and industry in Sweden and to what extent do these agreements address ownership, licensing, and benefit-sharing mechanisms for AI applications?
• What alternative models of data and IP governance (e.g., Bayh–Dole in the US, UK university IP frameworks, Dutch and Danish models) offer lessons for Sweden?
• What collaboration models (e.g., co-ownership, preferential access, revenue-sharing) can balance
  hospitals’ need for fair value with researchers’ incentives and open science principles and howcan such models ensure compliance with EU regulations while remaining practical for stakeholders?

Expected Effects

The project is expected to generate tangible benefits for hospitals, regions, an The project is expected to generate tangible benefits for hospitals, regions, and patients by creating fair and
transparent frameworks for data-driven AI innovation. By introducing legally sound benefit-sharing models, it will strengthen the region’s bargaining power, ensure returns on public investments, and provide faster access to AI applications for patient care. At the same time, it will foster sustainable collaboration ecosystems, build public trust through transparent agreements, and align open science principles with equitable value distribution. More concretely, the expected effects can be summarized as follows:

Increased Return on Public Investment – Benefit-sharing models will ensure that regions receive financial or in-kind returns from commercially successful AI tools.

Faster Access to Innovation – Hospitals will gain early or preferential access to applications
developed with their data, improving patient care.

Stronger Bargaining Position – Clearly defined rights will enhance regions’ leverage in negotiations with universities and industry.

Sustainable Collaboration – Aligning incentives across stakeholders will create long-term, mutually beneficial partnerships.

Improved Patient Trust – Transparent agreements will demonstrate that patient data directly
contributes to better care.

Open Science Alignment – Combining openness with contractual safeguards will balance
accessibility with fair value distribution.

Although piloted at Karolinska University Hospital, the project will serve as a prototype for other Swedish hospitals and directly support Region Stockholm’s priorities of strengthening innovation capacity, promoting value-based healthcare, and ensuring that public investments in health data generate maximum benefit for patients and society.