Allowed with disclosure. The funder lets applicants use AI and requires or asks them to say where they used it.
What the funder asks of applicants
- Researchers should identify any generative AI tools used in grant preparation and explain how the tools were used. The guidance is nonbinding, but disclosure can protect against research misconduct allegations. [1]
- Researchers should verify all results, including those from processes involving generative AI. [1]
- Researchers should check whether text or ideas from generative AI tools were copied from another person's work. They should cite those works appropriately. [1]
If an applicant breaks the rule
- Disclosure does not rule out a misconduct finding if the use departs from accepted practices of the relevant research community. Undisclosed or inaccurately described AI data generation or processing may be evidence of fabrication or falsification. [1]
- AI-generated fabricated citations (hallucinated references) in grant applications may not be fabricated data or results. In some cases, such as literature reviews, they could be the subject of misconduct allegations. [1]
- Institutions decide whether use of AI-generated text or ideas is plagiarism. Plagiarism means taking another person's ideas, processes, results, or words without appropriate credit. [1]
Rules for reviewers
Agency for Health Care Research and Quality has published no rules for reviewers.
Where this comes from
The funder's own words, from its own pages.
Researchers should disclose any generative AI tools used during research and manuscript and grant preparation.
Researchers should identify any generative AI tools used during their research and manuscript and grant preparation and explain how these tools were used.
However, disclosing generative AI tool use does not necessarily preclude a research misconduct finding.
Similarly, undisclosed or inaccurately described data generation or processing with AI tools may be evidence of data fabrication or falsification.
Researchers should verify all results, including those derived from processes involving generative AI.
Therefore, researchers should determine whether text or ideas they use from generative AI tools have been copied from another person’s work and cite those works appropriately.
Fabricated citations include those generated by AI tools but not reflecting real publications, which are often called hallucinated references.
Institutions must determine whether a respondent’s use of AI-generated text or ideas constitutes plagiarism.
This is a reading of what the funder has published, and it is not legal advice. A solicitation can set its own rule, so read the funding announcement before a PI applies.