Does anyone at your nonprofit actually know what your AI is doing?

What duty does a nonprofit board have to oversee AI?

A board's existing fiduciary duties, care and loyalty, now explicitly cover AI. Gene Takagi, one of the most cited voices in nonprofit law, laid this out plainly in a piece published August 16, 2026. The duty of care means a board must actively understand how AI is being used inside the organization, not wait for staff to bring it up. The duty of loyalty means weighing AI's efficiency gains against any harm it could cause to the mission, to beneficiaries, to staff, or to donor trust.

That is a real shift. Until recently, "we're experimenting with AI" was a staff-level update, mentioned in passing if at all. Takagi's argument makes it a governance matter. A board member who never asked how AI touches the organization is not neutral on the question. They are behind on a duty they already hold.

Source: Gene Takagi, Governance and AI: Fiduciary Duties, Nonprofit Law Blog, published August 16, 2026.

How big is the gap between what donors expect and what nonprofits actually tell them?

Three out of four donors want to know when an organization uses AI. About one in four organizations actually tell them.

That is the finding from a survey the Blackbaud Institute and Edge Research published on June 24, 2026: 76% of donors say it matters that organizations disclose when and how they use AI, but only 26% of professionals say their own organization does this. The survey ran in March 2026 across 1,389 social-impact professionals and 1,034 donors. I checked the numbers against independent coverage from Candid and The NonProfit Times before treating them as solid. A related figure from the same report: 68% of donors want their personal data protected in any AI use, and only 36% of organizations report doing that well.

Source: Blackbaud Institute and Edge Research, Bridging the AI Effectiveness Gap, published June 24, 2026.

Read those two numbers next to each other. Donors are not hostile to AI. They are not asking nonprofits to stop using it. They are asking to be told. Most organizations are not telling them, mostly because most organizations have not decided, at a leadership level, what there is to tell.

Why does this gap exist?

Because AI adoption happens the same way most work happens in a nonprofit: informally, task by task, without ever routing through a decision about what the organization is actually trying to accomplish this month.

I see this pattern constantly, and it is not unique to AI. A grant writer starts using a tool. A program manager tries another. Each is a reasonable choice made by someone doing their job well. But nobody above them can currently answer a simple question in one sentence: what is our AI doing on our behalf, and why. That is not a technology problem. It is the same execution gap that shows up whenever strategy travels from leadership down through a real organization and loses meaning along the way. AI just made the gap visible, because now a board member or a donor can ask about it directly.

What should a nonprofit leader do about this, starting this month?

Close it with three small moves, not a longer policy. Name an owner. Tie the ownership to a goal. Prepare the one-sentence answer.

Most AI policies I have seen are long, general, and unread. Naming an owner does more. Pick one person, not a committee, who can say what AI tools are in use and why. Then tie that to something concrete, a goal or a monthly priority, instead of a standing "watch this" item that never resolves. Finally, prepare the sentence. If a board member or a major donor asked today how you use AI and what you do to protect their information, could your ED answer without checking with anyone first? If not, that is this month's work. It costs nothing, and it is more useful than a binder nobody reads.

This is the same discipline I think about for execution generally, which is why DeTask exists: DeTask turns your mission into goals and keeps your leaders, people, and AI aligned around the results that matter. A governance policy answers what should happen. This is about who can say, in one sentence, what actually did.

Your donors already assume someone at your organization is watching what your AI does with their information and your beneficiaries' trust. The honest move is to make that assumption true, one owner, one goal, one sentence at a time. The question your board isn't asking yet is coming. It would be better to have the answer ready before someone asks it for you.

FAQ

Does my nonprofit need a formal AI policy?

Not necessarily, and not before you have an owner. A short, clear line of accountability does more than a long policy nobody reads. Name who owns AI oversight and tie it to a real goal first, then formalize a policy once you know what you are actually governing.

What exactly do donors want disclosed about AI use?

Mainly when AI touches communications or decisions that involve them, and how their personal data is protected inside that use. In the Blackbaud Institute and Edge Research survey, 76% called disclosure important and 68% specifically wanted data protection addressed.

Is AI oversight a board responsibility or a staff responsibility?

Both, but not equally. Per Gene Takagi's nonprofit governance commentary, the board carries the fiduciary duty to actively understand AI use, not simply wait for staff to report it. Staff typically execute the work; the board is accountable for oversight.

What should never be handed to AI without a human decision-maker involved?

Based on current nonprofit governance guidance, employment decisions, anything substituting for legal advice, and any use touching sensitive donor or beneficiary data should keep a human in the loop.