Latest / The Nonprofit Podcast / Ep 216| How to Write a Nonprofit AI Policy & Close the Donor Trust Gap

Transcript

Ep 216| How to Write a Nonprofit AI Policy & Close the Donor Trust Gap · The Nonprofit Podcast

• 0:00
  SPEAKER_01: Every nonprofit team is stretched thin right now.
• 0:04
  SPEAKER_01: Burnout, turnover, not enough hands for way too much work.
• 0:08
  SPEAKER_01: So when a tool shows up that can draft an appeal, answer an email, or summarize a lengthy report in three seconds, of course people reach for it.
• 0:17
  SPEAKER_01: At the same time, daughters are getting weary of exactly that.
• 0:21
  SPEAKER_01: They want to know their money is going somewhere real, not into a generic letter a machine wrote without a human ever touching it.
• 0:30
  SPEAKER_01: And that's the tension nobody is talking about out loud.
• 0:33
  SPEAKER_01: So today I want to talk about it.
• 0:36
  SPEAKER_01: The Nonprofit Podcast, powered by DonorBox.
• 0:40
  SPEAKER_01: Welcome to the Nonprofit Podcast.
• 0:42
  SPEAKER_01: I'm Jenna Lynch, head of community engagement here at DonorBox.
• 0:46
  SPEAKER_01: Each week we bring you practical strategies, clear insights, and inspiration to help you raise more, reach further, and take your nonprofit to the next level.
• 0:56
  SPEAKER_01: Here's what the data actually shows.
• 0:59
  SPEAKER_01: Charitable giving grew about 4.3% last year, which sounds pretty healthy on paper, but that growth was not even at all when you really dig into it.
• 1:10
  SPEAKER_01: Big organizations grew nearly 12%, while small community-based nonprofits actually shrank by more than 6%.
• 1:19
  SPEAKER_01: Some of that gap comes down to something surprisingly specific: how well an organization manages its tech tools, including artificial intelligence.
• 1:31
  SPEAKER_01: Roughly one in 10 nonprofits use AI in a genuinely governed way with a clear policy, clear ownership, clear boundaries.
• 1:41
  SPEAKER_01: Everyone else is mostly improvising, right?
• 1:44
  SPEAKER_01: Some daily use, but uh wildly different results.
• 1:49
  SPEAKER_01: And so that's the backdrop.
• 1:51
  SPEAKER_01: Now, here's the part that I think should make everyone sit up.
• 1:55
  SPEAKER_01: There are three numbers here that I want you to hang on to.
• 2:00
  SPEAKER_01: 76% of donors say organizations should tell them when AI was used to write to them.
• 2:08
  SPEAKER_01: 26% of organizations actually do that.
• 2:12
  SPEAKER_01: And only 14% have a written policy on AI use at all.
• 2:18
  SPEAKER_01: I want you to sit with that for a second.
• 2:20
  SPEAKER_01: That is not a small gap.
• 2:22
  SPEAKER_01: That is a 50-point trust gap.
• 2:26
  SPEAKER_01: And that is sitting quietly in the middle of our sector, just waiting for someone to notice.
• 2:32
  SPEAKER_01: And donors do notice, smart donors really notice.
• 2:37
  SPEAKER_01: They feel that gap when an email lacks transparency, and they can certainly tell when a thank you note reads like no human actually wrote it.
• 2:46
  SPEAKER_01: I don't think it's really about the technology itself.
• 2:50
  SPEAKER_01: Nobody is mad that a spreadsheet got built faster.
• 2:54
  SPEAKER_01: What people respond to, good or bad, is whether a human being was actually paying attention.
• 3:02
  SPEAKER_01: I like to think of this like a car with a really good autopilot feature.
• 3:09
  SPEAKER_01: Uh, autopilot is very popular.
• 3:12
  SPEAKER_01: It's great right up until the driver checks out completely, right?
• 3:17
  SPEAKER_01: The moment that nobody has a hand near the wheel, everyone in the car, and probably on the road too, feels it.
• 3:25
  SPEAKER_01: And that can be scary even before anything goes wrong.
• 3:29
  SPEAKER_01: I think in the same way, AI as a co-pilot can be useful, maybe even necessary at this point.
• 3:38
  SPEAKER_01: But if AI is the only one that's driving, I think that's where trust can break down for sure.
• 3:47
  SPEAKER_01: At the end of the day, AI should never be where your idea starts.
• 3:52
  SPEAKER_01: It should never be where your letter ends, your boots on the ground experience, your real relationship with your donors, and your organization's unique mission are what drive that strategy.
• 4:04
  SPEAKER_01: AI should just be the assistant to help you type a little faster and be a little more efficient.
• 4:10
  SPEAKER_01: I think it's worth being straightforward about something else, too.
• 4:15
  SPEAKER_01: I don't think that this is only a donor trust conversation.
• 4:19
  SPEAKER_01: Some of the sharpest critics in our space are not worried about donor perception at all.
• 4:24
  SPEAKER_01: They're worried about what's underneath the tool.
• 4:28
  SPEAKER_01: You know, the data centers running these massive models consume serious amounts of energy and water, often built near communities that already carry more than their share of environmental burdens.
• 4:40
  SPEAKER_01: Plus, a lot of the unglamorous work of training these systems is done by low-paid workers, often in the global south, without much support for what that work involves.
• 4:50
  SPEAKER_01: And then on top of that, we have to remember that AI will not fix structural bias on its own.
• 4:56
  SPEAKER_01: Large language models learn from historical data.
• 5:00
  SPEAKER_01: If past fundraising centers specific donor demographics or transactional habits, an unguided AI will just automate those exact same blind spots.
• 5:11
  SPEAKER_01: And the tool doesn't know your community's nuances, right?
• 5:14
  SPEAKER_01: You do.
• 5:14
  SPEAKER_01: None of that shows up in a donor letter, but it's real.
• 5:19
  SPEAKER_01: And a sector built on ethics does not get to just look away.
• 5:25
  SPEAKER_01: You know, data isn't just code or numbers, it definitely represents real people and real lived experiences.
• 5:32
  SPEAKER_01: And when we let an algorithm flatten a story or a donor relationship into some sort of fast prompt, we sacrifice data dignity for speed.
• 5:42
  SPEAKER_01: So that's something to consider, right?
• 5:44
  SPEAKER_01: Efficiency should never come at the expense of humanity.
• 5:48
  SPEAKER_01: So I think here the question becomes how do we navigate this responsibly?
• 5:55
  SPEAKER_01: Here are four steps that I've identified that I think can help you there.
• 6:01
  SPEAKER_01: So the first one is write down an AI policy.
• 6:07
  SPEAKER_01: It does not need to be 10 pages, it just needs to exist and someone needs to own it.
• 6:13
  SPEAKER_01: And this policy will be different for every organization based on your size, how it's been historically used, who your community is.
• 6:21
  SPEAKER_01: Uh so get a policy down.
• 6:24
  SPEAKER_01: It can be very simple and make sure that you're sharing it throughout your organization.
• 6:30
  SPEAKER_01: Two, a human should check everything before it goes out if it's used AI, every appeal, every report, every stewardship letter, you know, just actual human eyes reading through and doing a quality check.
• 6:47
  SPEAKER_01: Three, tell people if AI was used to create a report, say so.
• 6:55
  SPEAKER_01: It costs almost nothing to put that in a footer note.
• 6:59
  SPEAKER_01: And it definitely closes most of that 50-point trust gap by itself.
• 7:04
  SPEAKER_01: Be transparent.
• 7:05
  SPEAKER_01: Donors really appreciate that transparency.
• 7:08
  SPEAKER_01: And lastly, four, ask your vendors smart questions.
• 7:14
  SPEAKER_01: Don't be dazzled by the hype or paralyzed by the fear.
• 7:19
  SPEAKER_01: Ask your AI tech providers who built this model?
• 7:22
  SPEAKER_01: How is our data protected?
• 7:24
  SPEAKER_01: Where does our data go?
• 7:25
  SPEAKER_01: Is it being used to train any models?
• 7:28
  SPEAKER_01: Uh, you know, where does human sign-off happen?
• 7:32
  SPEAKER_01: Think about those questions that you need answers to and ask them.
• 7:36
  SPEAKER_01: Reach out and ask them.
• 7:38
  SPEAKER_01: You know, ethics does not stop at your own front door.
• 7:41
  SPEAKER_01: It continues into the vendors that you're using as well.
• 7:43
  SPEAKER_01: Now, this episode is not at all a call to throw out your AI tools.
• 7:49
  SPEAKER_01: It's just a call to keep your hand on the wheel.
• 7:52
  SPEAKER_01: Uh, I think AI can be incredibly useful.
• 7:56
  SPEAKER_01: Um, and this is actually the whole idea behind how DonorBox builds AI tools, including JAI.
• 8:04
  SPEAKER_01: You know, J it drafts, it flags, it suggests, it pulls reports, but it never sends anything to a donor on its own, right?
• 8:14
  SPEAKER_01: You're still a real person that reviews, signs off, and carries out the sending.
• 8:20
  SPEAKER_01: And just as importantly with JAI, your data stays your data.
• 8:25
  SPEAKER_01: When nonprofits use this, you know, their sensitive donor information isn't being used to train public models.
• 8:33
  SPEAKER_01: Everything's walled off.
• 8:34
  SPEAKER_01: Your donor history, financial records, private communications, they remain private, right?
• 8:39
  SPEAKER_01: Encrypted, compliant, and strictly used to help your organization raise more, not feed into a broader machine.
• 8:47
  SPEAKER_01: Um, so that's the model worth aiming for with any tool that you're using, wherever your organization is on the spectrum.
• 8:55
  SPEAKER_01: So it's not AI or no AI, but simply who's still driving and who owns it, right?
• 9:03
  SPEAKER_01: Uh DonorBox has a ready-to-use nonprofit AI policy template if you want to learn more.
• 9:09
  SPEAKER_01: And this includes a comprehensive breakdown of responsible AI use on our blog.
• 9:15
  SPEAKER_01: So I'll link both of those in the show notes.
• 9:19
  SPEAKER_01: If this got you thinking about where your own organization stands, start with one simple step.
• 9:26
  SPEAKER_01: Write down what your policy actually is, even if it's just a couple paragraphs.
• 9:31
  SPEAKER_01: That is the whole trust gap.
• 9:33
  SPEAKER_01: Starting to close one sentence at a time.
• 9:37
  SPEAKER_01: Thank you for being a part of the nonprofit podcast community.
• 9:40
  SPEAKER_01: Your work matters, and we are honored to support it.
• 9:43
  SPEAKER_01: If you enjoyed this episode, share it, leave a review, or subscribe to our YouTube channel.
• 9:48
  SPEAKER_01: Until next time, take that first step and stay inspired.
• 9:54
  SPEAKER_00: That warm feeling when you help someone, it's not just happiness, it's fulfillment.
• 9:59
  SPEAKER_00: And we believe it should be available to everyone, from frontline heroes to first time fundraisers.
• 10:05
  SPEAKER_00: Our tools empower you to help others.
• 10:08
  SPEAKER_00: This is our mission.
• 10:09
  SPEAKER_00: This is DonorBox, helping you help others.