A funder put $1 billion on AI. How a small nonprofit keeps its mission steady when the money moves

Most small nonprofits will never apply to the Gates Foundation. They still feel it when the Gates Foundation changes direction, because large funders help set what the rest of the field pays attention to.

This month the direction changed in public.

What happened

On September 14, the Gates Foundation announced plans to spend at least $1 billion over the next two years to expand access to AI. The foundation’s release splits it roughly 40% to education, 40% to health care, 10% to agriculture and 10% to building the digital foundation underneath.

The pushback came quickly. On September 17, The Chronicle of Philanthropy reported experts’ concerns about the education share, about $400 million. A political science professor pointed out that the foundation has not said who will evaluate the AI tools. An education author warned that spreading AI into under-resourced schools in the name of equity could backfire if it undercuts learning instead.

On September 23, the Chronicle reported how the foundation had shut down its Economic Mobility and Opportunity program, which had put about $500 million into anti-poverty work since 2016, and joined other philanthropists in backing an investment group, NextLadder Ventures, as it turned toward AI. Nisha Patel, who led a predecessor to the Gates anti-poverty program, said the work needs “patient capital.” The Chronicle’s account puts the contrast in plain words: there is no quick technological fix.

None of this says the pledge is wrong. It says that even the largest funder in the field is being questioned on evidence, scale and who benefits. A small nonprofit applying for AI-themed money will face the same questions, only with less room for error.

The pull, and why it is risky

When money moves, organizations bend toward it. A tutoring program adds “AI-powered” to its name. A food bank proposes a chatbot it has no staff to maintain. The grant arrives, the tool gets built, and two years later the organization runs a program it never meant to run, funded by a line that may not renew.

The alternative is not to ignore AI funding. It is to decide, before you write a word, whether the money serves what you already do.

A mission-fit test before you apply

Answer these five questions in writing. If you cannot answer yes to the first three, do not apply.

1. Would we do this work without the grant? If the AI part exists only because the funder asked for it, the program will not outlast the funding.
2. Can we name who benefits, and how we would know? Pick one measure you already track, such as time to a case plan, and state how it should change.
3. Do the people we serve have a say? If the tool touches clients, ask a few of them first. One expert the Chronicle quoted worried about side effects for the very students the money is meant to help.
4. Can we run it after the grant ends? Count the staff hours and subscription costs for year three, not year one.
5. Is the data safe to use this way? Client and donor records need a clear rule about which tools may see them and who can access the output.

A program that passes all five is one you can pitch with a straight face. One that fails the first question is a mission drift with a budget attached.

Writing the application honestly

If you do apply, three habits keep the proposal credible:

• Describe the tool as a means. Lead with the outcome for people, then explain where AI helps and where a person still decides.
• Show your evidence, even if it is small. A three-week staff pilot with a before-and-after time count is more convincing than a promise of transformation.
• Name the risks you will watch. Privacy, accuracy, and who gets left out if the tool assumes a smartphone or reliable internet.

Funders set their own rules on AI in applications, and they differ. Read each funder’s guidance, and remember your organization stays responsible for every claim in the proposal. This is general education, not fundraising-compliance or legal advice.

What to do this week

• List the funders you expect to approach in the next twelve months and note which ones have mentioned AI in their guidance or calls.
• Run the five-question test on any AI idea already in your pipeline, and share the answers with your board chair.
• Pick one internal measure you would use to show impact, and start recording it now, before any tool arrives.
• Draft a one-paragraph statement of how your organization uses AI today. If a funder asks, a clear answer is easier to write before the deadline than during it.

The money will keep moving. The mission is the part you decide to keep still.