AI Ethics and Prospect Research: Stop Asking “Should I?” and Start Asking “How?”

Research in Fundraising, , , , ,

The ethical debate around A.I. isn’t new. Here’s why that’s actually good news for prospect researchers.

Let me ask you something that might make you a little uncomfortable.

When you’re evaluating whether to use A.I. in your prospect research practice, what question are you actually asking?

If it’s “Is A.I. ethical?” — you’re asking the wrong question.

Not because ethics don’t matter. They matter enormously. Ethics is the linchpin of donor trust, and we ignore that at great peril. But the question “Is A.I. ethical?” is a dead end. It’s a question designed — consciously or not — to keep you frozen.

The question that actually moves you forward is: “How do I use A.I. ethically?”

And here’s the thing: you already know how to answer that. You’ve been doing it for years.

We Have Been Here Before

Consider these three concerns that have been raised about an information technology tool:

  • Ranking algorithms full of bias
  • Privacy and data harvesting concerns
  • Aggressive monetization practices

Are these complaints from 2024 about A.I.? Yes. Are they also complaints from 2003 about Google? Also yes.

We have been here before. And we made choices then, just as we’re making choices now.

Some people boycotted Google. That didn’t stop widespread adoption. Some people pretended not to see the wave coming — and got swept away in the current. And some people developed guardrails for responsible use and got on with their work.

I know which group I want to be in. And I’m guessing you do too.

Three Priorities That Matter Most

At Aspire Research Group, we don’t ask “Is this tool ethical?” in the abstract. We explore three priority items for every tool we adopt or update. Here’s how we think through each one.

Data Privacy

Every time we adopt a new tool — or an existing tool releases a significant update — we examine how data is stored, transmitted, and used. Then we make a deliberate decision about the level of risk we’re willing to accept. You might have an IT department or Advancement Services head that monitors this for you.

Concrete example: We use Google for search and maintain Google accounts for client work using our company emails. Our firm culture is to delete all browser cache daily to reduce search bias. We’ve also purchased accounts with Anthropic’s Claude and Microsoft’s CoPilot. The very first thing we did after purchasing? Investigated the privacy settings. Because the default — and I want you to hear this clearly — the default is almost always to share your data with the vendor. We changed that at the enterprise level immediately.

This is not revolutionary ethical reasoning. This is just being an informed consumer of technology. But it requires slowing down to actually do it, not just meaning to.

Data Sources

Knowing where your tool’s information comes from and equally important, what it doesn’t include, is fundamental to using it well.

A tool we’ve been excited about at Aspire is DonorAtlas, which uses generative A.I. to create detailed profiles. It’s genuinely impressive. But DonorAtlas draws exclusively from publicly available internet sources. It doesn’t include many public datasets, aggregated data behind paywalls, or sources that aren’t open to web crawlers.

That’s not a criticism — it’s just information. And because we understand it, we use DonorAtlas alongside tools like Kindsight’s iWave and Lexis Nexis for Development Professionals to fill the gaps. We also know that DonorAtlas struggles with non-traditional professions and (obviously) people with a limited digital footprint.

Understanding your tools isn’t optional. It’s the baseline for using them responsibly.

Bias Mitigation

Here’s the one that keeps me intellectually honest: bias is a slippery fish, and it always has been.

We’ve been actively examining our tools and practices for bias at Aspire for more than a decade — and it remains a persistent challenge. Traditional wealth screenings in the U.S., for example, rely heavily on real estate data, which systematically pushes minorities down the wealth list even when they hold significant wealth in private businesses or private stockholdings.

Does that mean we refuse to use wealth screenings? No. It means we advise nonprofits to treat screening results as a starting point, not a verdict. Eyes-wide-open use of imperfect tools is still responsible use — as long as your eyes are actually open.

With A.I., the bias questions are still evolving. DonorAtlas does a notably good job surfacing occupation data, which is a strong wealth indicator that screenings often miss. Could combining tools help us identify prospects that each tool alone would overlook? We’re testing it. Time will tell.

Before You Can Evaluate Ethics, You Need These Two Things

I want to stop here and say something that often gets skipped in the A.I. ethics conversation.

You cannot evaluate whether any information technology is ethical for your practice until you have a firm grasp of two fundamentals:

What information actually matters for your fundraising purpose?

Not everything you can find is relevant. At the Prospect Research Institute, we teach the “5 Building Blocks of the Profile” precisely because the temptation to gather everything is real — and counterproductive. A clear sense of what you’re looking for keeps your research focused and your use of tools purposeful.

Is the information source accurate and credible?

A.I. introduces a few new nuances here — a link can be real, an article can exist, and it still might not confirm the claim being made. But the foundational credibility framework? The C.R.A.A.P. test, developed by California librarians, remains highly relevant. (You can test your own source-spotting skills with our free Solid Intel course at the Prospect Research Institute.)

Master these two fundamentals and evaluating A.I. tools becomes much more tractable. Skip them, and no amount of ethical reasoning will save you from bad outputs.

A Candid Case for Leaning In

Here’s where I’ll get direct with you.

Refusing to engage with A.I. on ethical grounds is itself an ethical choice, and not necessarily the right one. If A.I. tools can help you find prospects more efficiently, surface connections that would otherwise be missed, and deliver better research to the gift officers who need it — then not exploring those tools has consequences too.

Responsible, effective use of available resources is key to fundraising success. That’s not a justification for ignoring risk. It’s a reminder that paralysis has costs too.

The ethical path forward isn’t to opt out. It’s to opt in — carefully, deliberately, and with your existing frameworks leading the way.

You already have those frameworks. You’ve been applying them to Google, to wealth screenings, to electronic databases, to social media, to every imperfect tool that came before this one.

A.I. isn’t asking you to throw out your ethics. It’s asking you to apply them.

So: how will you?

Additional Resources