Ethics, Accuracy & Funder Disclosure
This may be the most important lesson in the course. AI gives you speed, but speed without integrity is dangerous in fundraising, where your entire operation runs on trust. A single fabricated statistic, a leaked donor record, or a misrepresentation to a funder can cost a nonprofit its reputation and its funding. The good news: a handful of clear habits keep you safe. Master them and you can use AI confidently, ethically, and without fear.
What You'll Learn
- How to catch and prevent AI hallucinations before they reach a funder
- The privacy rules for donor and beneficiary data
- Whether and how to disclose AI use to funders
- A practical ethics checklist for every AI-assisted document
Accuracy: Defeating Hallucinations
AI's most dangerous trait is that it is confidently wrong. It will produce a fake statistic, an invented foundation, or a non-existent study in flawless, authoritative prose. In grant writing, where credibility is everything, an unverified claim is a loaded weapon pointed at your reputation.
Build these verification habits into every project:
- Treat every fact as unverified until checked. Statistics, dates, funder histories, citations, legal and tax claims — all of it.
- Verify against primary sources. Government data, the funder's own site, peer-reviewed studies, your own records. Not "the AI said so."
- Be extra suspicious of specifics. Precise figures ("73.4% of...") and named sources are exactly what AI fabricates most convincingly.
- Ask the AI to show its uncertainty. A useful prompt: List every factual claim in this draft and rate your confidence in each. Mark anything I must independently verify. This surfaces the riskiest claims for you.
Remember: when a proposal contains a false statement, your organization is responsible — not the AI vendor. Verification is not optional busywork; it is professional self-protection.
Privacy: Protecting Donor and Beneficiary Data
Public AI tools may use what you type to improve their models, and data can be stored on external servers. That makes privacy a serious duty. Follow these rules:
- Never paste personally identifiable information (PII) — donor names, addresses, emails, donation amounts tied to names, Social Security numbers, payment details — into a public AI tool.
- Anonymize beneficiary stories. Strip names and identifying details before using AI to shape a story; add real names back only with consent, in your own final copy.
- Know your tools' settings. Some AI plans let you turn off training on your data or offer business/enterprise tiers with stronger privacy guarantees. Use them for sensitive work.
- Follow your organization's data policy and the law. Regulations like GDPR (Europe) and various U.S. state laws govern personal data. When in doubt, leave it out.
A simple test: before pasting anything, ask "Would I be comfortable if this text appeared on a public website?" If not, do not paste it.
Disclosure: Should You Tell Funders You Used AI?
This is an evolving area, and norms differ. Some guiding principles:
- Check the funder's rules. A growing number of funders and government programs now ask about AI use, and some require disclosure or restrict it. Always read the application terms; if they ask, answer honestly.
- AI as a tool is generally accepted. Using AI to draft, edit, and research — with your verification and judgment on top — is widely viewed like using spell-check or a research assistant. You generally do not need to flag routine assistance unless asked.
- Misrepresentation is never acceptable. What matters ethically is that the proposal is true, that the work and outcomes are real, and that you stand behind every claim. Passing off fabricated data or a fake story as real is the violation — not the use of AI itself.
- When unsure, ask the program officer. A quick, professional question about their AI policy signals integrity, not weakness.
Bias and Fairness
AI models can carry biases from their training data, sometimes producing language that stereotypes the communities you serve. Review AI output with a critical eye: does it portray beneficiaries with dignity and agency, or lean on deficit-based, pitying, or stereotyped framing? You are the cultural and ethical filter. Edit anything that would not honor the people behind your mission.
Your Pre-Submission Ethics Checklist
Before any AI-assisted document goes to a funder or donor, run this list:
- Facts: Have I verified every statistic, citation, and funder detail against a primary source?
- Privacy: Did I keep all PII and confidential data out of public AI tools?
- Truth: Is every claim, story, and number genuinely true and something I can stand behind?
- Dignity: Does the language respect the people we serve?
- Disclosure: Have I checked and complied with the funder's AI policy?
- Ownership: Have I read every word and do I take full responsibility for it?
If you can answer yes to all six, you are using AI the right way.
A Realistic Example
A development assistant, rushing before a deadline, pasted an AI-drafted needs statement straight into a portal. It contained a fabricated statistic the AI had invented. A sharp program officer recognized the number was wrong, and the foundation — which had funded the nonprofit for years — quietly declined and grew wary of future requests. One skipped verification step damaged a decade-long relationship. The checklist above exists precisely to prevent that five-minute mistake.
Key Takeaways
- AI is confidently wrong; treat every fact as unverified until checked against a primary source, especially precise figures and named sources
- Never paste donor or beneficiary PII into public AI tools, anonymize stories, and use privacy-protective settings for sensitive work
- Check each funder's AI policy and disclose when asked; the ethical line is truthfulness, not the use of AI itself
- Watch for bias — you are the filter that ensures language honors the people you serve
- Run the six-point ethics checklist (facts, privacy, truth, dignity, disclosure, ownership) before anything reaches a funder