Before Fundraisers Ask What AI Can Do, We Need to Ask What AI Might Repeat

Fundraisers have always been good at working under pressure. 

There is pressure to raise more money, retain donors, acquire new ones, report better, steward faster, segment smarter, and prove impact in language that moves people.

So, when it comes to AI, in most of my conversations with fundraisers I see both a sense of relief and a worry about not hurting their donor community. Mostly, they come back with the same question — “What can AI do for us in a safe, responsible way?”

And I understand that interest — a tool that can draft the appeal, summarize donor notes, create stewardship emails, suggest audience segments, analyze giving patterns, turn one campaign into ten variations… all of it can help a small team do the work of a much larger one.

It makes sense that many fundraisers are curious. It makes sense that many are experimenting. It makes sense that some are excited.

But here is the question I want us to also think about in these conversations:

What if it helps us repeat old patterns faster? What do we do then?

Fundraising already carries many inherited assumptions: who is considered a "good prospect"; whose generosity is recognized; which donors are cultivated deeply and which are treated as one-time transactions; whose stories are centered; and which communities are seen as "beneficiaries" but rarely as decision-makers, donors, advisors, or holders of power.

AI does not enter fundraising as a neutral helper. It enters systems that already have histories, habits, hierarchies, and blind spots.

For example,

• If our donor data has always overrepresented wealth, AI will learn from that.
• If our communications have always centered on certain donor identities, AI will learn from that.
• If our segmentation has always valued capacity over community connection, AI will learn from that.
• If our stewardship has always rewarded the most visible forms of giving, AI will learn from that.

And then, because the output looks polished, fast, and confident, we may mistake repetition for insight.

This is where fundraisers need to slow down — not because AI is bad, but because they understand their work is deeply relational.

A donor email is not just content. A stewardship plan is not just another workflow. A segmentation model for auto-generated communications is not just about efficiency.

These are all decisions about attention, care, trust, power, and belonging.

When we use AI in fundraising, we are not only asking, “Can this tool help us write faster?” We are also asking, “What version of donor relationships is this tool helping us build?”

For example, imagine a fundraiser using AI to personalize donor emails. A small team saves time. Donors receive more relevant messages. The organization becomes more consistent.

On the surface, this may seem harmless, but there are deeper questions underneath.

• What data is being used to personalize those messages?
• Did donors understand that their past giving, event attendance, demographic information, or engagement history might be used this way?
• Are staff reviewing the messages before they go out?
• Could the language accidentally make assumptions about a donor’s identity, wealth, motivations, or relationship to the cause?
• If a donor says the communication feels invasive, who is responsible for listening and responding?

Or imagine an organization using AI to identify major gift prospects. Again, this may seem practical.

But here, too, are questions we must ask of the AI:

• What does the tool define as “potential”?
• Does it rely mostly on wealth indicators? Does it miss donors who give through mutual aid, volunteer leadership, community influence, or collective networks?
• Does it reinforce the idea that the most “valuable” people are those with the most financial capacity? Does it narrow the fundraiser’s imagination of generosity?

AI may help us see patterns. But it may also hide the values behind those patterns.

At a time when fundraisers are asking better questions about donor power, community-centric fundraising, trust-based philanthropy, ethical storytelling, and the relationship between money and justice, building a relationship with AI should not become a shortcut around those questions.

It should make those questions more urgent.

Before adopting AI in fundraising, we might ask:

• Where did this data come from?
• Who is missing from it?
• Would our donors, community members, or staff be surprised by how this information is being used?
• Can someone challenge or correct an AI-assisted decision?
• Are we using AI to deepen relationships, or only to increase output?
• Are staff being given the tools, training, and time they need, or are they simply being asked to do more with less?
• Who benefits if this works well?
• Who carries the risk if it does not?

These are not abstract technology questions. They are fundraising leadership questions — especially when AI can shift risk downward. 

A development manager may be told to “just use ChatGPT” without guidance, paid tools, privacy protections, or time to learn. Or a small nonprofit may rely on free tools because enterprise systems are too expensive. Or a fundraiser may feel expected to use AI because everyone else seems to be doing it, even if no one has clarified what is appropriate, ethical, or safe.

In that environment, AI does not reduce the burden. It redistributes the burden…often to the staff with the least power to question it.

And thinking about this kind of responsible AI use is especially important for small and mid-sized nonprofits. Larger organizations may have paid tools, legal review, IT teams, data staff, and governance committees. Smaller organizations may have one fundraiser, one database, one Canva login, and a board member asking why they are “not using AI yet.”

The question is not only whether AI is useful. It is whether the conditions around AI are fair.

• Are we funding the tool but not the training?
• Are we asking staff to experiment but not protecting them from mistakes?
• Are we encouraging innovation but not creating accountability?
• Are we expecting fundraisers to personalize more without asking what healthy personalization should look like?

None of us can, should, or even need, to become AI experts overnight. But we do need to become more fluent in the questions that protect trust.

Trust is the real currency of our work here.

Not data. Not dashboards. Not prompts. Not automation. Trust.

And trust is easy to damage when people feel watched, sorted, targeted, misrepresented, or reduced to a score.

The opportunity before us is not to reject AI. The opportunity is to refuse shallow adoption.

We can use AI to draft, summarize, brainstorm, analyze, and support our work. But we can also pause long enough to ask whether the work is becoming more human, more accountable, more transparent, and more just.

AI will not fix our bias. That is our job — to recognize, acknowledge, name, and work on the societal biases that show up in our data and operations.

For fundraisers, this may be the most important invitation: not to use AI because it is new, and not to avoid it because it is uncomfortable, but to approach it with care.

To ask better questions before scaling faster practices.
To protect relationships from becoming transactions.
To ensure the people affected by our systems are not invisible within them.

And if AI is going to become part of that work, then fundraisers deserve to help shape the conditions under which it is used — now, in the ongoing conversations.

As part of the third cycle AI Equity Project, I am currently gathering perspectives from nonprofit professionals, including fundraisers, on how AI is showing up in our sector, where it is helping, where it is creating pressure, and what responsible adoption should require. If these questions resonate with your work, I joyfully invite you to share your curiosity in this project and add your experience to the research (the report comes out in the fall).

I will also be teaching the AFP Deep Dive: The AI Advancement Lab in July for nonprofit professionals who want a practical, values-centered space to think through AI use, readiness, risks, and policies with more care. If you want a space to work through these questions together, registration is now open.