Published April 3, 2025, via Research News
On March 28, 2025, the National Institutes of Health (NIH) posted NOT-OD-25-081: Protecting Human Genomic Data when Developing Generative Artificial Intelligence Tools and Applications.
This notice indicates that while artificial intelligence (AI) tools and applications are significantly advancing biomedical research, the NIH encourages the research community to stay mindful of the potential risks of unintentional data disclosures when sharing AI tools and applications.
Specifically, NIH reminds researchers that:
- Users of NIH controlled-access data cannot use that data to train generative AI models without approval from NIH, per the Genomic Data Sharing (GDS) Policy and the Data Use Certification (DUC) Agreement.
- Generative AI models, including their parameters, developed with NIH controlled-access data must not be retained after the project is closed.
- Sharing controlled-access data with public generative AI tools (e.g. third party tools) via prompts or other user interfaces is prohibited.
For additional information on using controlled-access data responsibly, see the principles described in Using Genomic Data Responsibly Under the NIH Genomic Data Sharing Policy and the AI in Research: Policy Considerations and Guidance.
Sincerely,
Mark E. Lowe, MD, PhD
Vice Chancellor for Research
Associate Dean for Research, School of Medicine
Harvey R. Colten Professor of Pediatric Science
Visit the OVCR website for the latest information. Updates are posted frequently.