Nonprofit AI disclosure: what to tell donors and funders
November 16, 2026 · 4 min read
A staff member uses an AI tool to tidy a donor update. A grant writer uses one for a first draft of the needs statement. Neither mentions it, because nobody has said whether to. Nonprofit AI disclosure is the written answer to a question people are already deciding alone: when do we tell the person on the other end?
You don't need a long document. You need a decision, and a sentence people can use.
What Candid reports about nonprofit AI disclosure
On July 21, 2026, Candid's insights blog published a summary by Kyoko Uchida of the Blackbaud Institute report Bridging the AI Effectiveness Gap. Per Candid's summary, over three-quarters of donors say it's important to disclose when and how AI is used. About a quarter of nonprofit staff say their organizations do so.
The summary doesn't give sample sizes or survey dates, so read the numbers as a direction and not a measurement of your own supporters. The direction is still useful. Donors who care about this are more common than nonprofits who tell them.
Decide where the line is
Disclosure doesn't mean labeling every sentence. Sort your uses into three groups and decide each once.
Internal and invisible. Summarizing meeting notes, cleaning up a spreadsheet, drafting an agenda. These can usually go unlabeled, with the rule that a person checks the result.
Outward but edited by a person. A donor update, a newsletter, a grant narrative. Here a short statement is worth considering, and some funders may ask. If a funder's guidelines mention AI, follow them. If they don't, asking before drafting is the safer course.
Outward and unedited. A chatbot on your site, or an AI-generated image of a person. These should be labeled every time, and an image of a person in your program should usually be a real photo with consent.
Say a small animal shelter writes its appeals by hand but uses AI to draft social captions. The shelter decides captions are group two, adds a line to its policy, and moves on. That decision took 10 minutes.
Write the statement before anyone needs it
One or two sentences can cover most cases. For example: "We sometimes use AI tools to help draft and edit our writing. A staff member reviews everything before it's sent, and we never put donor information into these tools."
The last clause matters more than the first. What goes into the tool is a bigger question than what comes out, and it's covered in donor data and AI tools. Don't promise something your staff can't do. If someone has pasted donor names into a free chatbot, fix that before you say otherwise in public.
Put the statement where the policy lives, and make it part of what you give a new hire. If you have a written position, a nonprofit AI policy is the right home for it.
Settle who answers when someone asks
The board chair or the executive director will sooner or later hear from a donor, "Did a machine write this?" Decide in advance who answers, and give them the sentence above. A true, calm answer builds more trust than a careful non-answer.
Also decide what you will do if you get it wrong. If a staff member used a tool outside the rules, the response is to correct the practice and say what changed, in the same plain way.
The AI readiness assessment asks board members and staff the same questions about AI, in private, and shows where their answers differ. It takes about seven minutes each and is free, with no paid version. It doesn't tell you what to disclose. That's a decision for your board and your director.
A donor who learns how you use AI from you is in a different position from one who learns it from someone else.