Is it ethical to use AI on our donors without telling them?

It's Friday. Four o'clock. A development director sits down to clear her thank-you queue before the weekend. The AI assistant her organization subscribed to last quarter has drafted twelve notes — pulled from past correspondence, mimicked her tone, used each donor's first name, referenced the program each one gives to. The director edits two of them. Queues them all. Leaves for the weekend.

The donors get the notes on Monday. They feel seen.

None of them know AI helped write what they read.

Is that okay?

I'm asking because almost every organization I talk to is doing some version of this. ChatGPT in the development office. Claude in the comms team. A "personalization" tool somebody bought on a corporate card and never mentioned to the board. Donor data being pasted into prompts. Donor stories summarized into "case studies" by a system that may or may not be retaining them.

Most of it isn't malicious. Most of it is a tired development director trying to clear a queue on a Friday afternoon.

The question nobody has really asked yet — including us, mostly — is whether the donor would have agreed if we'd told her.

Donors don't see your AI. They feel your stance.

Donors aren't pulling back over AI yet. Most of them don't know to think about it. But there's a broader trust recession underway — they're giving more cautiously, asking more questions, watching the sector more carefully than they used to. The day a donor finds out exactly what's behind her donor-care touches is the day the trust recession lands specifically on us. That day is coming. The organizations that made the choice on purpose will be fine. The ones that didn't will spend a long time explaining.

Two paths.

Path one. Use AI to do more. Run more touchpoints, more personalization at volume, more appeals drafted, more meetings booked, more emails sent. More work done, same approach to the work itself. The donor relationship goes through the same machine it always did, just faster and at higher volume.

Path two. Use AI to do less. Hand the AI the repetitive work that was eating your development director's day — the data hygiene, the meeting notes, the templated acknowledgments that always needed editing anyway. The work nobody got into fundraising to do. Free her up for the thing the AI can't do — a real conversation with the donor whose mother just died, who isn't sure she's renewing this year, who needs five quiet minutes with the only person at the organization who knows her name.

The first path uses AI to scale fundraising. The second path uses AI to protect fundraising — to keep human relationships at the center while machines do the work that was never supposed to be relational in the first place.

Neither path is automatically wrong. But the choice gets made one way or the other the day the tool gets installed. Whether you make it on purpose is up to you.

Two questions worth holding before you implement anything.

Is it ethical to use AI on our donors without telling them we're using it?

And: is the AI we're adding doing more of what was never supposed to be ours — or doing more of what was always supposed to be ours?

These aren't compliance questions. There's no policy that resolves them for you. They're stewardship questions. And the answer your organization gives — even by default — is the answer your donors will eventually feel.

So a question for you, ED to ED:

If your largest donor asked you tomorrow whether AI was involved in the last thank-you she received from your organization, what's your answer? And have you decided yet what you want it to be?

The trust recession is real, even if AI isn't yet what's driving it. The patterns making nonprofits vulnerable to it — burnt-out staff, scarcity thinking, donor attrition you can't quite trace — usually predate the AI question. Our free organizational diagnostic names them by category in 15 minutes. If you want a place to work through these questions alongside other nonprofit leaders doing the same thing, Brazen Connections is free, every other Wednesday, and built for exactly this.