36 - Navigating Ethical Terrain: IRB 101 in the AI Era

Background:  The advancement of Artificial Intelligence (AI) technologies poses unique ethical challenges for Institutional Review Boards (IRBs) tasked with overseeing research involving AI. As AI research proliferates, IRBs must adapt to effectively address ethical concerns in this rapidly evolving field.

Program

Description:  This programmatic abstract presents our institutional approach to IRB management in the AI era to be better equipped for AI reviews. We developed and implemented a specifically tailored form (Form TS-AI) to help guide researchers and IRB reviewers to determine if FDA regulations apply for software as a medical device. The form incorporates a series of questions to help triage review type. The next phases of AI initiatives include updates to the device policy and developing draft guidance for researchers to guide them on providing the IRB with sufficient information to allow the IRB to make appropriate determinations regarding the use of AI. The IRB is adding IRB Members with experience in AI to better equip the IRB with expertise and skills to navigate conceptional, operational, and ethical considerations in AI studies. Finally, a checklist is being drafted to help IRB members make determinations about AI research.

Program

Assessment:  Preliminary surveys of IRB Expedited reviewers and research community were positive. IRB reviewers expressed better understanding of AI review regulations following the use of the form. The reviewers reported feeling more confident when evaluating AI research protocols and identifying potential ethical implications. Additionally, the program facilitated productive discussions among IRB members, leading to enhanced collaboration and decision-making processes.

Limitations: While Form TS-AI demonstrated positive outcomes, certain limitations were encountered during implementation. Factors such as resource constraints (i.e., available guidance) and varying levels of prior knowledge among IRB members may have influenced the form's effectiveness. Despite these limitations, the form is showing signs of success in fostering a culture of ethical awareness and proactive engagement within our institution's IRB for studies involving AI.

Discussion:  Moving forward, we propose to expand on the Form and add AI-specific training modules into research ethics curriculum to ensure ongoing competency in AI ethics among IRB members and the research community. Additionally, we advocate for the dissemination of our program model to other institutions facing similar challenges in AI research oversight. By sharing our experiences and best practices, we aim to contribute to the advancement of ethical standards in AI research on a broader scale.