In short
FDA has published no AI rule of its own on using AI to write or review grant applications. The only guidance that reaches FDA applicants comes from the Office of Research Integrity (ORI) at HHS. ORI recommends that applicants identify any generative AI tools used in preparing a grant application and explain how the tools were used. The guidance is a nonbinding recommendation. It applies to applications for Public Health Service (PHS) funding for biomedical or behavioral research. Disclosed AI use that departs from the accepted practices of the relevant research community may still be considered evidence of research misconduct. Applicants should check each FDA NOFO for added rules.
Permits AI. Applicants may use AI and stay responsible for what they submit.
For applicants
- FDA has published no AI rule of its own. For research applications, ORI recommends that applicants identify any generative AI tools used in preparing the grant application and explain how the tools were used. This is nonbinding guidance. It relates to the HHS research misconduct rule (42 CFR Part 93), which covers applications for FDA research funding. [1][2]
- Disclosed AI use that departs from the accepted practices of the relevant research community may still be considered evidence of research misconduct. AI data generation or processing that is undisclosed or inaccurately described may be evidence of data fabrication or falsification. [1]
For reviewers
Food and Drug Administration has published no rules for reviewers.
Sources
Researchers should identify any generative AI tools used during their research and manuscript and grant preparation and explain how these tools were used.
This guidance document is provided by the Office of Research Integrity (ORI) to assist entities that apply for or receive Public Health Service (PHS) funding for biomedical or behavioral research
This guidance document does not create or confer rights for or on any person and does not operate to bind ORI, the Department of Health and Human Services, or the public.
If a researcher discloses use of a generative AI tool in a manner inconsistent with the accepted practices of the relevant research community, that use may still be considered evidence of research misconduct.
undisclosed or inaccurately described data generation or processing with AI tools may be evidence of data fabrication or falsification.
the Centers for Disease Control and Prevention, the Food and Drug Administration, the Health Resources and Services Administration, the Indian Health Service, the National Institutes of Health, the Substance Abuse and Mental Health Services Administration, and any other components of HHS designated or established as components of the Public Health Service.
Every extramural or intramural institution that applies for or receives PHS support for biomedical or behavioral research, biomedical or behavioral research training, or activities related to that research or research training must comply with this part.
This page summarizes what the funder has published and is not legal advice. Policies change, and one solicitation can add its own rules. Confirm against the funding announcement before you apply.