AI policy template for funders
A funder’s AI policy tells applicants whether they may use AI tools on an application, what they must disclose, and how your team reviews what they send. Keep it to one page and post it with your guidelines. A published answer saves every applicant from guessing and saves your staff from answering the same question again and again.
In a 2024 Candid survey, two in three foundations were still undecided about AI-generated applications. Applicants are already using these tools, and many now learn them at workshops funders run. A short policy gives every applicant the same answer.
Free to copy and adapt. Square brackets mark what you replace. Nothing here is tied to a particular funder. Read each funder’s guidelines for required headings, page limits, and attachments before you paste a finished draft into their application form.
The template
Purpose and scope
Say what the policy covers.
This policy explains how applicants may use AI tools when preparing applications, reports, and other materials for [foundation name].
What applicants may use AI for
Name the uses you accept. Applicants follow a list more readily than a principle.
Applicants may use AI tools to research funding, organize their materials, draft and edit text, and check an application against our requirements. [Add any condition you set.]
Applicant responsibility
Put the accountability on the applicant, where it already sits for any draft.
The applicant is responsible for every statement in the application. Facts, figures, and citations must be accurate and traceable to a source. A named person at the applicant organization approves the final text.
Disclosure
Decide whether you want a statement. If you do, keep it short and say it does not affect scoring.
[Choose one] We do not ask applicants to disclose AI use. / Applicants describe in one or two sentences how they used AI tools, if at all. Disclosure does not affect scoring.
How we review
Detection software is unreliable. OpenAI withdrew its own detector in 2023 for low accuracy, so tell applicants you judge the work itself.
Reviewers score every application on our published criteria. We do not use AI-detection software to screen applications.
Confidentiality in review
Hold your own team to the standard you ask of applicants.
Staff and outside reviewers do not enter applicant materials into AI tools unless [approval and approved tools].
Support for applicants
If you teach AI to grantees or give them tools, say so here.
We offer [workshops, office hours, or sponsored tools] to help applicants use AI well. Details: [link].
Ownership and review
Give the policy an owner and a date.
Policy owner: [role]
Last reviewed: [date]
Next review: [date, or sooner if practice changes]
Skip the blank page
Quillify drafts each application from the applicant’s own documents and holds it to your published requirements, which makes a policy like this one simple for grantees to follow. Funders can also sponsor Quillify seats for their grantees and see how the group is doing, while each organization’s documents stay its own.
Questions people ask
Should a foundation publish an AI policy?
Yes, if you want every applicant working from the same rules. A one-page policy posted with your guidelines answers the question once for everyone.
Can funders detect AI-written applications?
Not reliably. OpenAI withdrew its own AI text detector in 2023 because of its low accuracy, and a 2023 Stanford study found that detectors flagged most essays by non-native English speakers as AI-generated. Scoring applications on your published criteria is the fair test.
Related guides: How funders treat AI in grant applications. How to review an AI-drafted grant proposal: a checklist. What to say when a funder asks how you used AI. How to write your organization’s AI policy for grant work. All free templates.
Let Quillify write the first draft.
Browsing and keyword search cost nothing. There’s no card to enter.
