You've started using an AI tool for application drafts, and then you wonder whether the funder is fine with it. There's no single rule. One funder may limit it, another may ask you to describe how you used it, and the next may say nothing at all.
This article shows you what some funders have published, where to look for the rule that applies to you, how to disclose AI use when asked and what to do when the answer isn't written down.
Policies differ, so look at each funder
Funder policies on AI fall into three broad groups, and the same organization can meet all three in one season.
- Limits on what you may submit. The funder says AI-developed content isn't acceptable, in whole or in part.
- Disclosure expectations. AI use is allowed, and you're asked or encouraged to say how you used it.
- Silence. The notice doesn't mention AI at all.
Because policies change, treat anything below as an example of what funders write, and read the current text for the funder you're applying to.
What NIH says
The National Institutes of Health published notice NOT-OD-25-132 on July 17, 2025. It applies to applications submitted for the September 25, 2025 receipt date and later. In it, NIH says applications that are either substantially developed by AI, or that contain sections substantially developed by AI, are not considered the original ideas of the applicant. This covers whole applications and individual sections.
The same notice says AI may be appropriate to assist in preparation for limited aspects or in specific circumstances. It warns that AI use can lead to plagiarism or fabricated citations. NIH also says it will use technology to detect AI-generated content, and that if AI use is found after an award, it may refer the matter to its Office of Research Integrity and take enforcement action such as disallowing costs or withholding future awards.
The notice also caps applications at six per principal investigator or program director per calendar year, which matters if you were planning to use AI to produce more applications.
NIH treats its reviewers differently. Under NOT-OD-23-149 (June 23, 2023), reviewers may not use generative AI tools to analyze applications or write critiques, or upload application content to them, because it can't be known where the data is sent, stored or viewed, or how it may be used later.
What NSF says
The National Science Foundation's notice on generative AI, dated December 14, 2023, takes a different approach for proposers. It encourages them to say in the project description the extent to which, if any, generative AI was used and how it was used to develop the proposal. It also states that proposers are responsible for the accuracy and authenticity of their submission. For reviewers, NSF prohibits uploading proposal content or review information to generative AI tools that NSF has not approved.
What a private funder can look like
A published policy from a foundation shows a third pattern. The Wenner-Gren Foundation says use of generative AI is not prohibited, that responsibility for the originality and accuracy of content rests entirely with the applicant, and that applicants are encouraged to disclose use in a confidential section of the application. It also says that uploading application materials to a generative AI platform is strictly forbidden for its reviewers.
Notice what the three funders share. None of them treats AI as an excuse for errors, and each one protects what applicants send in by restricting reviewers.
Where to check
For each funder on your calendar, look in this order:
- The notice or guidelines. Search the document for "artificial intelligence", "generative", "AI" and "originality". Check the application instructions and any attachment templates, not only the main notice.
- The funder's website. Look for a policy page, FAQ or grants-management terms. Foundations often put these outside the guidelines.
- The portal. Some application forms include a question or a field about AI use.
- The program officer. If you can't find a rule, ask. Talking to a program officer covers how to make that call useful.
When disclosure is requested
If a funder asks or encourages you to disclose, do it plainly. Say what the tool was used for, what you did with the output and who is responsible for it. Never leave out AI use where disclosure is required, and never try to hide it.
Say a small youth-arts nonprofit is applying to a funder that invites disclosure. A suitable statement might read: "We used a generative AI tool to produce a first draft of the organizational background section from our annual report and to check section consistency. Our program director revised the draft and verified all figures against our records." It is short, specific and puts a named role in charge of accuracy.
When the notice is silent
Silence isn't a rule in either direction. It means the funder hasn't written one that you can find. Before you rely on that, check the funder's general terms and conditions, and ask the program officer one specific question: "We use an AI tool to help draft and edit our applications, and a staff member reviews and verifies everything. Is there anything you'd like us to note?"
Keep the reply. If the funder says to go ahead, record the name and date in your log. If it asks for disclosure, include it. Either way, the draft you submit still has to be accurate, original to your organization and checked by a person, because those expectations show up in every policy above.
Keep a policy log
Make a simple table with one row per funder, with columns for the funder, the source you found, the date you checked, what it says and what you decided. When a funder's notice changes from one cycle to the next, you'll see it. When someone asks why you used a tool on one application and not another, you'll have the answer.
What to hand off to AI and what to keep can help you decide what the log should allow in your own process. How to read a funding notice helps you find these passages quickly.
Your next step
List every funder on your calendar for the next three months. For each, search the notice and the funder's site for an AI policy and record what you find in the log. For any you can't resolve, write a two-line question to the program officer.
When a funder permits AI assistance, SoliContraX gives you a draft to edit and own, built from your organization's materials. You decide what to keep and what to disclose. See how to get started.
Sources
- NOT-OD-25-132, Supporting Fairness and Originality in NIH Research Applications (NIH)
- NOT-OD-23-149, The Use of Generative Artificial Intelligence Technologies is Prohibited for the NIH Peer Review Process (NIH)
- Notice to Research Community: Use of Generative Artificial Intelligence Technology in the NSF Merit Review Process (NSF)
- The Wenner-Gren Foundation Generative AI Policy (Wenner-Gren Foundation)