Aaron Barbour, one of our consultants, has just completed a Masters in charity finance and management at the Centre for Charity Effectiveness, Bayes Business School. Here’s a short blog he wrote for one of his assignments about grantmaking, written in a more academic style, to a restricted word count.
I love techno. And jungle, hip-hop, rock and jazz too. Lately it has been Ambrose Akinmusire’s trumpet on repeat, that searching, haunting sound that never quite goes where you expect. What I love is that each is made by people using the instruments of their day: the acoustic guitar, then the electric, then the sampler, and now software that writes a whole track from a line of text.
The kit keeps changing but what the music is ‘for’ never does. It still has to move us. Techno knows this. It is, after all, the sound of machines, built on synthesisers and drum machines by people who hear music in the technology itself. Which is roughly where grantmaking finds itself today. A fourth industrial revolution has put AI into the hands of foundations (Schwab, 2016), and many are unsure whether to play it, let alone how.
So I want to borrow one genre in particular, jazz, to explore this. Of all of them, jazz is the one that knows what to do when there is no score in front of you, which is where philanthropy has found itself since ChatGPT 3.5 broke through in November 2022.
There is no score
You cannot read the score for a technology while it is still being written. David Collingridge (1980) called this the dilemma of control. Early on, you can shape a technology but do not yet understand its effects. By the time you understand them, it is too embedded to change. Wait for certainty and you forfeit any power to influence it. Act early and you risk acting on a problem you only half understand. There is no safe option, only a choice about which risk to take. And philanthropy, if history has anything to teach us, should be prepared to take risks (Davies, 2026) with such a fast paced, influential technology.
Yet the stakes are not merely operational efficiency and productivity. As the Center for Effective Philanthropy argues, AI is not only changing how foundations work, it is changing the world that philanthropy is trying to change (Forster and Iyer, 2026). Meanwhile, the band has already started playing. 88% of charities now use AI tools in their day to day work, up from 61% just two years ago (Charity Digital Skills Report, 2026). Around 60% of foundation staff are experimenting with AI, yet most grantmakers have no strategy or plan for how to use it (Chronicle of Philanthropy, 2026). We are at the equivalent stage of a 5-year learning to play the violin. Painful! As the technologist Rachel Coldicutt puts it, FOMO is not a strategy (Careful Industries, 2025), but nor is standing still and hoping it passes.
So the band plays on. But over what?
The changes you play over
People think jazz is the absence of rules. Musicians know the opposite is true. Improvisation isn’t freedom from structure, it’s freedom over one (Weick, 1998). Donald Schön (1983), who used jazz to explain how good professionals actually work, called it ‘reflection in action’. Not winging it, but thinking on your feet over years spent honing your craft. For a grantmaker, the key a piece is played in is grounded in its values, governance and mission. Ignore this and improvisation becomes a free for all, though some free jazz enthusiasts may disagree.
Some of that structure is being composed. Floridi and Cowls (2019) offer five ethical principles for AI in society (beneficence, non-maleficence, autonomy, justice and explicability). A neat structure to solo over. Closer to home, the Charity Commission (2024) advises that trustees remain responsible for the decisions a machine helps them reach. A human must stay in the loop.
You can already hear this working in practice. The Patrick J. McGovern Foundation’s Grant Guardian, now used by 400+ funders in the USA, applies AI to help financial due diligence but it does not make funding decisions. It gives the grants officer a clearer picture and leaves the judgement with the human (McGovern Foundation, 2025).
Conversely it also means knowing when not to reach for the software. Meredith Broussard (2023) calls the opposite instinct ‘technochauvinism’, the assumption that technology must be ‘the’ answer. Often, though, the right response to a grantee isn’t a better algorithm. It’s simply a phone call.
Playing with the band
No one improvises alone. You play by listening and answering. And this is where philanthropy is most out of tune. 90% of non-profit leaders say they want to use AI more, but only 17% say a funder has ever talked to them about it (Center for Effective Philanthropy, 2025). That is a soloist playing over the room instead of with it.
The better players are learning to listen. In May 2026, the National Lottery Community Fund launched a £3m programme, with UK Community Foundations and the tech charity CAST, to fund communities to shape AI rather than simply use it (NLCF, 2026). It is supporting 50 community organisations to build locally-rooted AI tools and flag when an algorithm is failing the people they serve. It is the funder’s version of listening by paying for communities to make their own sound, rather than mastering the track for them.
Like any new sound, AI in grantmaking will splinter into sub-genres, one for due diligence, another for strategy, another for grantee support. Whoever leads each one early will quietly set its rules for those who follow.
Coda
The greatest solos are never the freest. John Coltrane could pour out the notes, but what made him so sublime was that he was listening to the band, the room, the song underneath. No finished score has been written for this fourth age of AI, and waiting for one is really just a decision to let someone else set the stage. The grantmaker’s job is one any decent musician has always had: to play, play through the changes, with the band and never stop listening.
If there is one practical note to end on, it is this. Faced with Collingridge’s dilemma, shape the technology early or understand it once it is too late, the rational response is not to wait, nor to rush, but to fund adaptively. Because nobody yet knows quite how this AI ‘music’ will go. The most useful thing a funder can do is offer flexible, unrestricted funding to give charities the room to compose and improvise.
Techno proved a machine could make music but only because a human first heard the music in it. AI is grantmaking’s new machine. It will play nothing worth hearing until funders compose and play with the communities and charities they support.
REFRENCES
Amar, Z., 2026. Early insights from the 2026 Charity Digital Skills Report survey. Zoe Amar Digital. Available at: https://zoeamar.com/2026/03/26/early-insights-from-the-2026-charity-digital-skills-report-survey-ai/
Broussard, M., 2023. More than a Glitch: Confronting Race, Gender, and Ability Bias in Tech. Cambridge, MA: MIT Press.
Careful Industries, 2025. Five lessons for careful AI adoption. Blog post from Careful Industries. 14 February 2025 https://www.careful.industries/blog/2025-2-five-lessons-for-careful-ai-adoption
Center for Effective Philanthropy, 2025. AI With Purpose: How foundations and nonprofits are thinking about and using artificial intelligence. Cambridge, MA: CEP.
Charity Commission for England and Wales, 2024. Charities and artificial intelligence. Available at: https://charitycommission.blog.gov.uk/2024/04/02/charities-and-artificial-intelligence/
Chronicle of Philanthropy, 2026. Can AI make grant seeking easier and grant making more refined? Available at: https://www.philanthropy.com/news/can-ai-make-grant-seeking-easier-and-grant-making-more-refined/
Collingridge, D., 1980. The Social Control of Technology. London: Frances Pinter.
Davies, R., 2026. SMM781 Principles and Management of Philanthropy, Grantmaking and Social Investment, ‘Public Good by Private Means: Philanthropy policymaking in historical perspective’ (lecture). Bayes Business School, City St George’s, University of London. 22 May 2026.
Floridi, L. and Cowls, J., 2019. A unified framework of five principles for AI in society, Harvard Data Science Review, 1(1).
Forster, C. and Iyer, L., 2026. AI Is reshaping the world – foundations must revisit their strategy. Center for Effective Philanthropy. Available at: https://cep.org/blog/ai-is-reshaping-the-world-foundations-must-revisit-their-strategy/
McGovern Foundation, 2025. Grant Guardian. Available at: https://www.mcgovern.org/our-work/data-solutions/grant-guardian/
National Lottery Community Fund, 2026. £3m to help communities shape the future of AI and confront the ‘wisdom gap’. Available at: https://www.tnlcommunityfund.org.uk/news/3m-to-help-communities-shape-the-future-of-ai-and-confront-the-wisdom-gap
Schön, D. A., 1983. The Reflective Practitioner: How Professionals Think in Action. New York: Basic Books.
Schwab, K., 2016. The Fourth Industrial Revolution. Geneva: World Economic Forum. https://www.weforum.org/stories/2016/01/the-fourth-industrial-revolution-what-it-means-and-how-to-respond/
Weick, K. E., 1998. Improvisation as a mindset for organizational analysis, Organization Science, 9(5), pp. 543–555.
