Using AI Ethically (and Moderately) in Fundraising

A field guide to the practical basics in using AI responsibly in nonprofit fundraising—without losing your donors’ trust or your own voice

Twenty-five years ago, my “AI” was a filing cabinet, a stack of Chronicle of Philanthropy clippings, and a very patient program officer who let me call with follow-up questions. Today, I use AI as an assistant and a productivity hack.

I have a healthy skepticism of AI. I’m not at all a fan of AI energy and water use. Or nefarious developments like AI autonomous weapons systems and the pretty much certainty of AI falling into roque hands in the future. The world needs guardrails, agreements, and all kinds of solutions to environmental and human impacts.

And, and, and.

But there are also many benefits to AI, like breakthroughs in disease. The productivity hacks are real. And as an independent freelance writer and fundraiser, it’s either learn to master it or get left behind.

So I use it, including for this newsletter. I enjoy mastering the prompts, tools, and techniques and becoming more proficient. I always do a check and human hand edit before I hit send or publish. I’ve seen enough tech trends and evolution to know that novelty and wisdom are not the same thing.

That’s the heart of ethical AI fundraising: using these tools with the same rigor you’d bring to a major, six-figure proposal package. Enthusiastically, but not carelessly.

A journey through the canals of Venice, Italy © Tonya Hennessey

The Organizational Policy Gap That Needs Fixing

Here’s the uncomfortable part. Most of our sector is still improvising when it comes to AI use. Seventy-six percent of nonprofits still don’t have a formal AI policy, according to Nonprofit Tech for Good’s State of AI in Nonprofits: 2025 report.

This means three out of four organizations are letting staff use generative tools with no shared guardrails on what’s appropriate, what’s disclosed, and what’s off-limits. That’s not a responsible AI nonprofit strategy. That’s improv night, and your donor list is not the audience you want workshopping material on.

My PREMIUM piece this Thursday delves into creating your organization’s AI Policy.

Foundations Are Watching—and Some Are Saying No

Candid’s August 2024 survey of grantmakers found that 23% of foundations will not accept grant applications containing AI-generated content, while only 10% explicitly welcome it. My caveat is that this is two- year-old data that surely has evolved.

Many funders are still figuring out where they stand. One funder put it plainly in that same survey: “Our due diligence requires reviews of financials and site visits where AI impact would be limited or neutralized.”

Translation: funders aren’t anti-AI, they’re anti-shortcut. AI proposal writing ethics is as much about whether you used a tool, as whether the finished proposal still sounds like your organization, has a human filter, and holds up under scrutiny and due diligence.

My Three Non-Negotiables

After decades of raising grant and gift dollars the analog way, here’s where I’ve landed on AI ethics for nonprofits:

Donor data stays out of the AI tool. Full stop. This is my hardest line. Your donor database holds names, giving histories, wealth indicators, sometimes health or family details—deeply personal, confidential information that donors entrusted to your organization, not to a third-party server. Feeding that into a public AI tool is a breach waiting to introduce itself. Keep AI use in the drafting and productivity room, not the donor database.

Say so, plainly. If AI helped shape a proposal, letter, or report, disclose it when it’s material to the relationship and the situation. Transparency isn’t just a legal hedge; it’s good manners, and donors and funders alike respect it rather than being surprised.

Watch for bias like you’d watch a campfire in dry brush. AI bias is real—these tools learn from historical data, and historical data reflects historical inequities. It’s an incomplete picture. If you’re using AI to screen prospects, draft case statements, or summarize program outcomes, read the output with a critical eye for language that flattens nuance, leans on stereotypes, and is incomplete.

Always use your human filter to factcheck and edit the final product. This goes for emails as much as documents. Your mission deserves better than autopilot.

The Groove, Not the Grind

I’m being transparent: I use AI to help draft this very newsletter—research, structure, a rough pass that rewrite in my own voice, checked against every fact I can verify. I lay out exactly how, and why, on my updated About page, under “How I Make It.” Check it out.

AI isn’t always the villain in this story, and it isn’t the hero either—it’s more like a competent research assistant and productivity multiplier with a chronic overconfidence problem. Use it moderately and appropriately, always with a human hand, disclose it honestly, and keep your donors’ data exactly where it belongs: with you, protected, and never connected to an AI tool.