IRB-draft-generator: A generative AI tool to streamline the creation of institutional review board applications

Abstract

The Institutional Review Board (IRB) is fundamental to conducting ethical research involving human subjects. IRB applications require detailed descriptions of the research and specific indications of how the research will be implemented. This can be difficult for inexperienced researchers. Preparing the application is a significant time commitment, even for experienced researchers. In order to lighten the administrative burden on busy clinical professionals, this software application will automatically generate a draft human subject research protocol (the most laborious element of an IRB application) based on responses to a short form. This technology uses generative AI and a custom literature search plug-in to draft the protocol from succinct, user-provided details. User inputs include a brief description of the research, including the hypothesis, inclusion/exclusion criteria, and the study design type (e.g., randomized clinical trial). This tool can expedite the IRB application creation process, provide additional consistency for reviewers, and may reduce clinician researcher burnout through a reduction in clerical work thereby facilitating participation in meaningful research.

Keywords

• Institutional review board
• Generative AI
• Large language models