Ask a nonprofit leader whether their team could use more help, and the answer comes before you finish the question. With tight budgets and lean staff, AI looks like welcome breathing room for decisions and campaign resources. But, then ask what happens to supporter data inside those tools, and the answer turns cautious.
If you lead a nonprofit, that tension probably sounds familiar. You want the capacity AI promises, but you’re unsure how these tools handle information and where their answers come from. Meanwhile, corporate partners, donors, and volunteers have started asking whether your organization has an AI policy.
Nonprofits don’t need every answer about AI yet, but they do need to start asking the right questions. Here are five worth putting on your leadership agenda.
1. What are our staff and volunteers putting into AI tools right now?
Start with what is actually happening. Many AI tools are free, open in a browser tab, and useful within minutes: a development associate drafts an appeal, a volunteer coordinator summarizes survey responses. In fact, a 2026 survey of 917 nonprofit staff and executives by NTEN and The Bridgespan Group found that exactly this was happening: 53% of respondents reported shadow use, meaning informal AI use outside organizational guidance. Executives reported it more often than staff, at 57% versus 49%.
That doesn’t point to carelessness. Lean teams reach for whatever helps them keep up. But supporter names, giving histories, and volunteer details belong to people who trusted your organization with them, and that trust travels with the data.
A useful first step is a candid, blame-free conversation with staff and volunteers. Ask which tools they use, for which tasks, and what kinds of information go in. You can’t set a direction until you know your starting point.
2. Do we know what happens to our information once it enters the tool?
With a CRM, you generally expect your data to be secure, and vendors are typically clear in their documentation and agreements about what data is and isn’t shared.
General-purpose AI tools, including large language models like ChatGPT and Claude, feel different to many leaders. There are more unknowns about what information gets shared and how it is handled.
That uncertainty makes many nonprofits nervous about putting their data into these AI tools, and the concern runs through the whole organization—and they’re not alone in that unease: In the NTEN and Bridgespan survey, 55% of executives named data privacy as a major barrier to AI adoption, and 56% of staff called it a significant concern.
You can narrow the unknowns by asking the same questions you would ask any vendor. Where does our data go? Who can see it? Is it kept, and for how long? Is it used to improve the tool? What controls do administrators have?
The answers can vary by tool, plan, and settings, so read each tool’s own terms rather than assuming. For anything involving personal supporter information, involve your IT advisor or legal counsel, who are better placed than we are to interpret them.
3. Where is the AI getting its information, and how do we check it?
Leaders who are enthusiastic about AI often have a second worry: the sources. AI can be genuinely helpful when you are weighing a decision or building campaign resources. But an answer that sounds confident can still be wrong, outdated, or impossible to trace back to a source.
For a nonprofit, the stakes are practical. An incorrect impact figure in an appeal, a misstated program detail in a sponsor deck, or a fabricated reference in a grant narrative can damage the credibility you have spent years building.
Decide in advance which kinds of output need a human to check them before anyone relies on them. Anything with facts, figures, or claims about your programs belongs on that list. A helpful mindset is to treat AI output as a first draft from a fast colleague who sometimes gets things wrong. The draft saves time, and a person still owns the final version.
4. If a donor, sponsor, or corporate partner asked about our AI policy tomorrow, could we answer?
Corporate volunteers, sponsors, and donors are increasingly concerned about their own data privacy, and they are asking nonprofits directly whether an AI policy exists. The question often arrives at a high-stakes moment, like a sponsorship renewal, a grant application, or a corporate volunteer day—and organizations that can share a policy quickly show they’re prepared.
However, most organizations would struggle to do that today. The NTEN and Bridgespan survey of nonprofit executives found that:
- 37% don’t currently train staff on how to use AI tools
- 58% don’t have an AI roadmap (only 8% do)
- 40% have written guidance in place for what data can be used with AI
You don’t need a perfect, forty-page document to begin. You need something that exists, that leadership has agreed on, and that you can hand over when asked. Just as important, someone should own it. Decide who answers these questions when they arrive, and who updates the answer as your use of AI changes.
5. Where could AI genuinely help our team, and what are we ready to try?
Small teams need every tool they can get, and AI can take real work off a lean team’s plate. But many have no dedicated budget for it. In fact, 57% of executives reported none.
With limited money and time, nonprofits have to be selective. To do that, ask where AI helps your team most, whether that’s drafting campaign materials, organizing information for a decision, or summarizing long documents, and start there. A narrow, well-chosen use is simpler to check for accuracy, simpler to keep free of supporter data, and simpler to explain to a partner who asks.
No matter what, plan to revisit your answers regularly, since the tools change quickly and so will your partners’ expectations.
Start with the questions
Few nonprofits have settled answers to all five of these questions, and that is a normal place to be. The leaders who will be best positioned are the ones who start asking now, so that decisions about AI are made on purpose and with their supporters’ trust in mind.
At Giveffect, we spend our days working with nonprofits on technology, supporter data, staff capacity, and day-to-day operations, and these questions come up in many of those conversations. We aren’t a legal or compliance resource, so treat this list as a way to start the discussion with your team, your board, and your counsel.
FAQs
Does my nonprofit need an AI policy?
Many nonprofits are finding that corporate partners, sponsors, and donors are starting to ask whether one exists, and a documented policy is becoming something funders and partners expect to see. Your counsel can advise on what yours should contain.
Is it safe to put supporter data into AI tools like ChatGPT or Claude?
There is no single answer, because what happens to information depends on the tool, the plan, and the settings. Read each tool’s terms, ask the vendor-style questions listed above, and involve your IT advisor or legal counsel before entering personal supporter information. Until you have answers, many organizations choose to keep identifiable supporter details out of general-purpose AI tools.
How can a small nonprofit use AI when staff time is limited?
Lean teams tend to find AI most helpful for decision-making and for building campaign resources. A sensible way to begin is with tasks that don’t involve personal supporter data, then widen the circle as your policy and your confidence grow.
Who should be responsible for AI decisions at a nonprofit?
Name one leader who owns the policy and the answers to partner questions, and bring in operations, development, technology, and board perspectives. Shared ownership helps the policy reflect how people really work.
How often should we revisit our approach to AI?
Plan to review it regularly, since tools and partner expectations are changing quickly. Reviewing it alongside other technology and data decisions can help it stay current.