Most nonprofits use AI, but few have rules for using it safely

Nearly every nonprofit has staff using AI, but very few have written rules for it. A 2025 AI Equity Project survey found that 92 percent of nonprofits use AI in some way. Only about 15 percent have an organizational policy for responsible use. A recent Nonprofit Quarterly essay argues that this gap is the real problem.

Use without a plan

Most nonprofits use AI on an ad hoc basis, meaning one person at a time with no shared tools or routines. Staff draft emails, clean up grant language and summarize meeting notes. The author says most of the value goes to those individual employees, not to the organization. Many groups also depend on one local AI enthusiast who answers everyone's questions.

Why no policy matters

Without a policy, the author warns, there are no controls to protect sensitive data. The essay also says that unguarded use is how bias gets repeated and trust gets worn down. Many executive directors, it says, assume they can deal with this later.

The fears are real

The author does not dismiss the worries. Those listed include environmental costs, job loss, made-up answers, bias in training data and less human connection. These concerns may hit hardest the communities nonprofits serve. The essay says the cause of the gap is not apathy. It points to overwhelm, legitimate fear and a sector-wide conversation that has been too abstract to be useful.

On jobs, the author says AI will change roles, but for small groups the bigger story is time. Picture a three-person organization where one person spends 12 hours a week on data entry, thank-you letters and report formatting. AI could hand much of that time back for relationship and strategy work.

Why a person must check the work

AI chatbots, called large language models, produce text by predicting which words are likely to come next. They do not look up facts. There is no separate fact-checking step, so an invented statistic reads just as smoothly as a true one.

The author's answer is a human review of anything AI helps write for the public, including grant narratives and impact reports. The motto is "first not final." Before anything goes out, check these against a source document:

• Names
• Dollar figures and dates
• Citations and quotes
• Anything about your own programs

What to do

Any team using AI, nonprofit or not, can start small. Write a short policy that says what data must never go into these tools. Name who reviews AI-assisted work before it goes public. The essay also raises energy use as a concern, but the text provided cuts off before it explains that point.