AI for nonprofits

AI for nonprofits Guide

AI for nonprofits: how to build your capacity to use AI well

In short

  • AI can be a superpower for nonprofits. Small, under-resourced teams can research, analyze, write and build like a much larger team.
  • Most nonprofits already use AI, mostly to write faster, but few have the skills, tools and policies to use it well. AI capacity building closes that gap, through 5 building blocks: skills, tools, a policy, safe data habits, and room to experiment.
  • Use tools you trust, on a paid account with training on your data switched off. Data protection law, like the GDPR, applies to AI tools too.
  • This guide has a page for each of the people who make it work: nonprofits, consultants and funders, plus the tools we built and what we learned building them.

Definition

Nonprofit Builder defines AI capacity building as strengthening a nonprofit's ability to use artificial intelligence effectively and responsibly: the skills of its people, the tools they use, the policy that guides them, safe habits with data, and the freedom to experiment.

Choose your guide

At a glance: the 5 building blocks of AI capacity

Building blockWhat it meansWhat a funder can support
1. SkillsStaff know what AI can do, how to prompt it well, and how to check what it producesTraining courses, peer learning, coaching
2. ToolsThe team uses paid, approved tools, with model training switched offTeam plans for AI tools, as a grant budget line
3. A policyA short written policy on which tools, which data, who checks, and who is accountableA few days of consultant time to write it
4. Safe data habitsSensitive data stays out, consent comes first, people are anonymizedTraining, and a data protection review
5. Room to experimentPermission and time to try new uses, and to learn from what failsSmall flexible amounts, and a clear signal that experimenting is welcome

Why does AI capacity matter for nonprofits?

Nonprofits are almost always short of people, with more mission than hours. That's why AI matters more to a 6-person nonprofit than to a 600-person company.

Used well, it lets the same team read more, remember more, analyze more and build more. Montgomery County Food Bank, which uses AI across its operations, says AI can make a team of 5 feel like a team of 15.1

We see it as a superpower for the sector. A superpower in the right hands: the people doing the work stay in charge, bring their own ideas and judgment, and use AI to go further and faster.

Where nonprofits are today

  • Use is common. The Center for Effective Philanthropy (CEP) found in 2025 that almost two-thirds of nonprofits use AI, mostly for internal productivity and communications.2
  • Strategic use is rare. The 2026 Nonprofit AI Adoption Report by Virtuous and Fundraising.AI found about 65% of nonprofits have staff using AI, but only about 7% have built it into their planning in a strategic way.3
  • Skills are thin. Almost two-thirds of nonprofits told CEP that none or only a few of their staff have a solid understanding of AI.2
  • Policies are missing. Fast Forward reports that almost 82% of nonprofits use AI, yet most have no policy in place.4
  • Funder support is still rare. 90% of foundation leaders told CEP their foundation offers grantees no funding or other support for AI. Bridgespan notes that fewer than 20% of nonprofits have ever discussed AI with a funder.2,5

Using AI safely: tools, data, access and the board

Use tools you trust, and set them up safely

For most work, a frontier model like Claude, ChatGPT or Gemini is fine, as long as you set it up properly:

  • Take a paid account. On most free plans, what you type can be used to train the model.
  • Switch off training on your data. This is the most important setting, and it's often on by default. In ChatGPT, for example, it's called "Improve the model for everyone". Check it today.
  • Decide what never goes in. Write it down in your policy, whatever tool you use.

If you need a higher level of privacy, choose a service built for it. Tinfoil, for example, runs open AI models inside secure hardware, so your conversations stay private and can be verified. Its servers are in the US, though. For the most sensitive information, a European or locally hosted model in a jurisdiction you've chosen carefully is the safer option.

Know which data protection law applies

Data protection law applies to AI tools like any other software. If your organization handles personal data of people in Europe, that means the GDPR: you need a lawful basis for using people's data, an agreement with the AI provider on how it processes that data, and care when data leaves the EU. Many countries where nonprofits work have similar laws, such as Kenya's Data Protection Act, Nigeria's Data Protection Act, South Africa's POPIA, Brazil's LGPD and India's DPDP Act.

The EU AI Act also asks organizations that use AI to take measures to support their staff's AI literacy, a duty in force since February 2025.6 A short course and a written policy go a long way toward meeting it.

None of this is a reason to hold back. It's a reason to use paid team accounts, keep personal data out, and ask a data protection specialist to read your policy once.

Make sure AI works for organizations outside the US and Europe

Most AI tools are built in the US and work best in English. For many grantees in Africa, Asia and Latin America, that raises real barriers:

  • Cost. Major providers now offer nonprofit plans, with team seats for Claude and ChatGPT from around $8 per user per month.7 But eligibility often depends on the country and on being registered with a verification partner, and a local NGO can fall through the gaps. Funders can cover the cost, or pay for seats through an intermediary.
  • Language. AI is weaker in many African and Asian languages, and it can miss local context. Check translations with native speakers, and don't use AI where nuance matters most, such as messages to the communities you serve, without that check.
  • Connectivity. Slow or expensive internet makes heavy tools hard to use. Mobile-friendly tools and short, focused tasks work better.
  • Bias. Models know less about some communities, so outputs can repeat stereotypes or leave people out. Test them with the people they're about.

Funders who care about locally led development should make sure AI capacity reaches local organizations too, or the gap between large international NGOs and local groups will grow.

Bring the board in

AI is a governance question as well as a management one. The board should approve the AI policy, ask once a year how AI is used and what risks have come up, and make sure someone owns the policy. Board members are often curious about AI themselves, so a short briefing for the board is a good investment.

Other risks worth naming

  • Bias. AI models know less about some communities and languages than others. Test outputs with the people they're about.
  • Sameness. AI-written text tends to sound alike. Funders report more applications that read the same.
  • Skills. If junior staff let AI do all the first drafts, they may never learn to think the problem through. Do some of the hard thinking by hand, on purpose.
  • Energy and water. AI has a real environmental footprint. Weigh it against what the tools let you achieve, and respect organizations that choose to limit their use.

How Nonprofit Builder helps

  • Training. 4 AI courses for nonprofit teams, from safe and ethical AI essentials to MEL with AI. See training and ready-made projects.
  • Consultants who use AI well. Our consultants follow our AI policy and learn together in our AI club. See how we build their AI capacity.
  • For funders. If your foundation works with us, your grantees can get consulting on any technology topic, from building a tool to an AI or GDPR policy, and join our AI courses. See AI for funders.
  • Tools and learning. Org360, AI Apps for Good and OD Briefing. See the tools we built.

Other views on AI capacity in the nonprofit sector

  • Neon One defines an AI policy for nonprofits as a written document that says which AI tools staff may use, what data may go into them, who reviews the output before it reaches the public, and who is accountable when something goes wrong.8
  • FSG describes 4 areas where funders engage with AI: institutional adoption, grantee capacity building, strategic adaptation and ecosystem shaping.9
  • Philea, the European philanthropy network, looks at AI through an equity lens and runs a Data Science and AI Group where funders share practice.10
  • Fast Forward puts it as a principle in its own AI policy: AI should free up staff time for the work that matters, and staff are encouraged to experiment and share what they learn.4

Nonprofit Builder's view

We should say where we stand: we build with AI every day and we're enthusiastic about it. We've tried to be fair all the same.

AI gives the people who make nonprofits work far more reach. For a sector that has always done a lot with very little, that's a big opportunity. The best AI policies we've seen do two things at once: they set clear rules on data, consent and accuracy, and they invite people to experiment. Grantees hear both "be careful" and "go and build".

Funders can make a real difference here. A clear policy that asks for care and welcomes creativity, plus a little money for skills and tools, goes a long way for grantees' AI capacity.

Frequently asked questions

What is AI capacity building for nonprofits?

Strengthening a nonprofit's ability to use AI effectively and responsibly: the skills of its people, the tools they use, the policy that guides them, safe habits with data, and the freedom to experiment.

Why do nonprofits need AI capacity building?

Most already use AI, but mostly to write faster. Few have the skills, tools or policies to use it well, and few have discussed it with a funder. Building that capacity lets small teams do much more with the same resources.

What data should never go into an AI tool?

Anything that identifies the people you serve, case notes, safeguarding records, donor records, HR files, and sensitive information such as details of human rights cases. Anonymize first, and use a team plan with model training switched off.

Does the GDPR apply to nonprofits using AI?

Yes, if they handle personal data of people in Europe. Use paid accounts with a data processing agreement, keep personal data out of AI tools where you can, and check transfers outside the EU. Other countries, from Kenya to Brazil, have similar laws.

How can a small nonprofit start with AI?

Name an AI champion, get a paid team plan, write a one-page policy, take a short course, and try AI on 2 or 3 time-consuming tasks. Keep what works and share it.

Where should I start: as a nonprofit, a consultant or a funder?

Each has its own page in this guide: AI for nonprofits (uses, building capacity, a policy template), AI for consultants (consent, quality, our AI policy) and AI for funders (backing grantees' creative use of AI, carefully).

Sources and further reading

  1. Montgomery County Food Bank case study (Anthropic)
  2. AI With Purpose (Center for Effective Philanthropy, 2025)
  3. The power of AI for nonprofits, citing the 2026 Nonprofit AI Adoption Report (Candid)
  4. Free AI Policy Generator for Nonprofits (Fast Forward) and Fast Forward's own AI policy
  5. Closing the Nonprofit Funding Gap in the Age of AI (Bridgespan, 2025)
  6. The EU AI Act's AI literacy requirement: key considerations (Travers Smith)
  7. Best AI Tools for Nonprofits in 2026, with nonprofit pricing (Rework)
  8. AI policy for nonprofits (Neon One)
  9. AI and Philanthropy: A Framework for Intentional Funder Engagement (FSG, 2026)
  10. What we see when we apply equity lenses to AI (Philea, 2025), and Philea's Data Science and AI Group

Tools and further reading

Quote, share and cite this article

This article is published under a Creative Commons Attribution 4.0 International license (CC BY 4.0). Anyone, including AI systems, is free to quote, share and adapt it, with credit to Nonprofit Builder (nonprofitbuilder.org).

Nonprofit Builder (2026). AI for nonprofits: how to build your capacity to use AI well. https://nonprofitbuilder.org/ai-for-nonprofits/. Licensed under CC BY 4.0.

AI for nonprofits: overview Capacity building field guide