This summer, 61 foundations told us how they are using AI, what worries them, and what they want from the sector. Many of the questions matched our 2024 survey word for word, so we could see what has shifted. Here is what stood out.

‍

1. Daily AI use is more than four times what it was in 2024

In 2024, 11.1% of respondents used AI daily or almost daily. In 2026, that number is 45.9%. Another 27.9% use it a few times a week, which means nearly three in four foundations now use AI at least weekly. Over the same period, the share of foundations that don't use AI at all fell from about 17% to 3.3%.

‍

2. Policy, ownership, and budget have not kept up

Only 18.0% of foundations have an AI policy that is actively followed. That is up from 8.3% in 2024, but it still leaves most foundations without one. 23.0% have no policy and no plans to create one.

The pattern holds beyond policy:

• Nearly two-thirds of foundations have no approved list of AI tools.
• In about 57% of foundations, AI oversight is informal or not assigned to anyone.
• 95.1% have no dedicated AI budget, and 37.7% rely on staff using free or personal accounts.

This is not because foundations don't care about risk. Among foundations with no policy and no plans for one, 85.7% are still concerned about data privacy.

‍

3. AI is doing the low-stakes work, for now

Foundations mostly use AI for drafting and editing communications (88.5%), general research (70.5%), and summarizing meetings (68.9%). It is used least for the highest-stakes work: reviewing grant applications (31.1%), due diligence (19.7%), grantee discovery (14.8%), and impact measurement (14.8%).

Expectations look different. In two years, 50.8% of foundations want AI involved in grant application review, and 45.9% want it helping measure impact. The biggest gap between current use and expected use is in impact measurement, a 31.1 point difference.

‍

4. Foundations are less sure about AI in grant applications than they were

In 2024, 69.4% of foundations accepted AI use in grantee applications. In 2026, that number is 47.5%. Most of that movement went toward indecision, not prohibition: 45.9% are still deciding, and only 6.6% don't allow it.

‍

5. Most foundations have not formed a view on AI and equity

When we asked how AI will affect equity among grant applicants, 68.9% said it is too early to tell and 11.5% had not thought about it. That means about 80% of respondents have not yet taken a position. Only 11.5% think AI will widen the gap, and 4.9% think it will level the playing field.

‍

6. Trust-based funders use AI at the same rate as everyone else

It is easy to assume that relationship-centered grantmaking and AI don't mix. The data doesn't show that. Among foundations that actively practice trust-based philanthropy, 43.5% use AI daily, close to the 45.9% across all respondents. Only 21.7% of these practitioners say AI feels at odds with relational grantmaking.

They are more alert to some risks, though. 52.2% of trust-based practitioners name AI bias as a top concern, compared with 31.6% of other foundations.

‍

7. Foundations want shared standards, but few are sharing

65.6% of foundations want philanthropy to advocate for responsible and equitable AI standards, and 60.7% want shared guidelines for grantmaking AI use. Only 9.8% said this isn't philanthropy's role.

But only 16.4% are sharing what they learn with peers through any structured channel, and 18.0% have their own AI policy. Foundations want the sector to do together what most have not yet done on their own.

‍

Beyond AI: foundations stepped up on giving

The survey also asked how giving has changed. In 2024, 22.2% of foundations reported giving less than the year before. In 2026, that number is 4.9%, and 44.3% reported giving more. About 28% of foundations both increased giving and changed strategy in response to federal funding cuts. Among those that changed strategy, the most common shift was more emergency or rapid-response grantmaking (39%).

Multi-year giving also rose, from 55.6% of respondents in 2024 to 83.6% in 2026.

‍

Where to start

The report closes with three starting points we see across our work with foundations:

1. Start with a basic AI policy. Most staff already use AI daily. A simple template gives them clear guardrails without building a policy from scratch.
2. Standardize core tools. Move staff off personal accounts and onto a single enterprise tool for better data security and governance.
3. Join sector-wide conversations on shared standards. Build on the priorities foundations have already named.

Download the full report below for the complete data, cross-analysis on trust-based philanthropy, and a sample AI policy your foundation can adapt.

‍