UW Research

September 2, 2025

For the Record- September 2, 2025: New Requirements for SOM AI Research, DCTI Exceptions

In this Issue:

New Requirements for School of Medicine Research Involving AI

Effective September 2, 2025

The Human Subjects Division (HSD) has collaborated with UW Medicine to develop a set of requirements to mitigate risks associated with research using artificial intelligence (AI). The intention of these requirements is to implement safeguards while minimizing the burden on researchers by augmenting the IRB review process rather than developing a separate workflow or review process for research involving AI. The requirements are designed to align with the new UW Medicine Policy on Use of Artificial Intelligence in the Healthcare Setting.

The scope of the requirements described in this announcement is limited to research reviewed by the UW IRB involving the use of AI that is:

  1. led by School of Medicine (SOM) Principal Investigators (PIs), AND,
  2. involves either UW Medicine Data or targeted enrollment of UW Medicine patients.

HSD will pilot the new process for 1 year, with a midpoint check-in, before we consider applying any aspects of this process to other schools and departments within the University of Washington.

What’s New?

  1. Expanded definition of Human Research.

HSD’s current interpretation of the regulatory definition of a human subject does not capture some research that should be reviewed by the IRB to mitigate risks to subjects that may result from re-identification of their data by AI systems. We have therefore revised our interpretation to include secondary use of de-identified data (including medical images) that are at high risk of re-identification and could lead to significant harm if re-identified. This means that  IRB review will be required for some research involving AI that did not previously require IRB review.

  1. New guidance for SOM research involving AI and required supplement.
The new Guidance for School of Medicine Research using AI provides a standardized and risk-based approach to the review of research involving AI. It provides a framework to help the IRB identify risks associated with AI, determine whether risks have been appropriately mitigated, and communicate the IRB’s expectations to researchers. The guidance applies to research reviewed by the UW IRB involving the development of AI systems and the use of AI as a tool to facilitate the administration of a research study (e.g., recruitment, safety monitoring, data analysis).
  • The guidance does not apply to the use of AI as a tool for the administration of research when: 1) the AI tool(s) have been approved for use in clinical care and UW Medicine business operations as described in UW Medicine’s policy and 2) are being used for their approved purpose.
  • The guidance does not apply to existing studies unless a modification is submitted to add a new use of AI.
  • The guidance does not apply to research reviewed by an external non-UW IRB.
For research covered by this guidance, the UW IRB now requires researchers to complete and submit the  Artificial Intelligence supplement form with their IRB application. The supplement is designed to be used with the guidance to develop and describe a plan to address risks associated with research involving the use of AI and to provide the IRB with the information necessary to complete its review of the risk mitigation plan.
  1. UW Medicine Security Review.

A security review will be required for research reviewed by the UW IRB and research reviewed by an external non-UW IRB when it involves using AI outside of a Secure UW Medicine Environment. Researchers will be prompted to complete the security review by questions in the Request External Review form and the AI supplement form. The security review will be completed in 3 to 5 business days and can happen concurrently or prior to IRB review.

Thank you. HSD is grateful to UW Medicine for their partnership on this initiative, the UW researchers who provided feedback on the guidance and supplement, and our friends at the DigiPen IRB for sharing with us their expertise in this arena.

Questions. Contact hsdinfo@uw.edu with any questions about the new requirements.

 

Exceptions to the Diversity in Clinical Trials Policy

Reminder

The UW Diversity in Clinical Trials (DCT) policy aims to promote inclusive research practices that reflect the diversity of the populations we serve. It comes with additional submission requirements (i.e., Diversity Plan) for new studies submitted to HSD after January 1, 2026. However, not all clinical trials are subject to the policy requirements (e.g., submission of a Diversity Plan). The policy allows for some exceptions including:

  1. Phase 1 or Earlier Clinical Trials
These early-stage trials are designed to minimize risk and assess safety using conservative eligibility criteria. Because variability in response is high and pre-clinical data are limited, certain populations may be justifiably excluded or underrepresented.
  1. Pilot and Feasibility Studies
The goal of pilot and feasibility work is not to test hypotheses about the effects of an intervention, but rather to assess the parameters or components of a larger scale study. Given the smaller scope of pilot and feasibility studies, they may not include the range of populations for which the treatment was intended.
  1. Clinical Trials Involving Small Populations
Some trials target populations so small that subgroup analysis is not feasible. Some examples include rare diseases, unique study populations (e.g., astronauts), isolated environments, emergency situations, restricted resources coupled with an urgent need, and tailored therapies. If the available target population is <100, or the total study enrollment is <30 across all sites, the population may be too small to allow for statistical calculation.

If you have questions about applicability of the DCT requirements for your research, contact hsdinfo@uw.edu.