Almost every nonprofit is using AI now. In one 2026 survey of 346 nonprofits, the 2026 Nonprofit AI Adoption Report from Virtuous and Fundraising.AI, 92% said they use AI in some way. Nearly half, 47%, said they have no AI policy at all. And 65% described their use as reactive and individual: one person trying a prompt here, another using a personal account there.
If that sounds like your office, you are not alone. And your staff are not doing anything wrong. They found a tool that saves them time, and they started using it. That is exactly what good employees do.
The problem is that without rules, every person decides on their own what is okay to paste in. A donor list to clean up. A client's story for the newsletter. Notes from a hard conversation with an employee. Nobody means any harm. Nobody told them where the line is.
Why a written rule beats a reminder
Many leaders have said something at a staff meeting. "Be careful with AI." "Don't put anything private in there." That is a good start, but a reminder fades. New hires miss it. Volunteers never hear it. And "be careful" means something different to everyone in the room.
A written policy gives everyone the same answer. Your board adopts it, so staff know it has backing. It tells a new hire on day one what is allowed. And it gives you something to point to when a funder, a member, or a board member asks how your organization uses AI.
There is one more benefit people don't expect. Tools like Claude and ChatGPT let you load your rules into the organization's account, so they apply to every conversation. At the nonprofit I lead, our rules are set once and apply every time. If someone starts to paste in something on the "never" list, the tool stops and reminds them. It is a helpful backstop, not a guarantee. People are still responsible for what they enter.
The six areas every AI policy needs
You don't need a 20-page document. You need clear answers in six areas, written so a new hire could follow them.
1. Confidentiality: what never goes in
Name the information that stays out of AI tools. For most nonprofits and associations, that list includes names and contact details of clients, members, and donors; health information; anything about children; donor giving and financial records; and personnel files. The simplest rule of all: when in doubt, leave it out.
2. Authority: draft or do?
Draw the line between preparing something and acting on it. Drafting a grant application is help. Submitting it is an action a person takes. AI can draft an email; a person sends it.
3. Professional limits: when a specialist weighs in
Decide what needs a qualified person. AI can summarize a contract, but it does not replace your attorney. It can explain a budget line, but your accountant signs off on the audit. If your mission touches health, legal, or financial services, say so plainly.
4. Voice and audience: who are we talking to?
Describe how your organization sounds and who reads your work. Plain language? Warm? Formal for funders, friendly for members? This is the part that saves you the most rewriting.
5. Human review: who approves it?
Spell out what a person checks before it goes out: public posts, emails to members or donors, grant applications, board materials, and anything that touches money. My favorite way to say it: AI tools draft. People decide.
6. Format and identity: the everyday details
How your organization is named, which templates to use, your default file format, your brand colors. Small things, but getting them right every time is what makes AI output feel like yours.
Keep it short
The best policies I have seen are short. Six clear sentences do more than three pages of principles, because people actually read them. Start with the basics, adopt it, and improve it once a year.
How to get started this month
Ask your staff which AI tools they already use, with no blame attached. You need an honest picture first.
Pick the tools you will approve, and use organization accounts, not personal ones.
Draft your policy using the six areas above.
Have someone who knows your data review it. If you handle health, financial, or children's information, ask your attorney to review it too.
Take it to your board for adoption, then walk your staff through it.
Load it into your AI tool's organization settings so it applies to every conversation.
Review it once a year, and any time you add a new tool.
Start with a free template
I built a free, fill-in-the-blank AI Use Policy Starter Template from the policy I use at the nonprofit I lead. It covers all six areas, includes a simple chart of what can and can't go into AI, and has a signature block for board adoption. Download it, make it yours, and take it to your next board meeting.
If you would rather have help setting up AI for your team, from the policy to custom prompts and staff training, that is what my Nonprofit AI Setup does. Either way, start with the rules. Everything else is easier after that.
Quick answers
Does a small nonprofit really need an AI policy?
Yes. Small teams often have the most to gain from AI and the least time to fix a mistake. A one- or two-page policy is enough to start.
Should we just ban AI instead?
A ban is hard to enforce, and it can push use onto personal accounts where the organization has no visibility. Clear rules and approved tools usually protect you better.
What should never go into an AI tool?
Names and contact details of the people you serve, health information, anything about children, donor financial records, personnel files, passwords, and legal matters. When in doubt, leave it out.
The template and this article are general information, not legal advice.