How We’re Using AI to Further a Power Shift

For better or worse (or somewhere in between), the future is now. AI isn’t coming—it’s here. And, despite some of the hype, it actually presents opportunities for funders to increase, rather than decrease, community involvement in grantmaking.

Since Camden Giving was set up in 2017, we’ve worked on the belief that people experiencing Camden’s inequalities should be the ones making decisions about how grants are awarded. Over the years, we’ve worked with more than 350 citizens in this way and have even supported national funders to move towards more participatory approaches. Participatory grantmaking has fostered social cohesion, strengthened community capacity, and changed lives in Camden. It’s a happy coincidence that it also leads to better funding decisions.

This blog isn’t about the ethics of AI (which absolutely need to be discussed). Instead, we want to share a practical example of how we’re using AI to strengthen community power.

Isn’t AI the Opposite of Participatory Grantmaking?

Community intelligence is what powers our grantmaking. Our grantmakers work in teams—first as a broad voting group of 350+ citizens who set grantmaking priorities, then in smaller teams of 10–12 citizens who make individual funding decisions. No one’s intelligence is valued over another’s, and we intentionally bring together people with different lived experiences to paint a rich picture of the solutions Camden needs.

Community knowledge is undervalued in academia, philanthropy, and the wider world, but its benefits are real and tangible. A particularly stark example I’ve seen was speaking with young people in the aftermath of a tragic murder in Camden. The way that kind of trauma affects their ability to travel around the borough is profound. AI and research alone won’t know that. They can’t make decisions that rebuild safety for communities. But there is a role for AI in supporting and amplifying community decision-making.

AI and research live in the past, making predictions based on what’s already happened. Participatory grantmakers are living in the present, in a rapidly changing world—and that’s crucial.

Using AI to Visualise Solutions

AI generated image of a homework club.

When we met with a group of young people, we gave them a very open brief: spend £100,000 on anything that supports young people experiencing crisis due to growing up in low-income households. They shared ideas based on what they’d seen work, but no one had exactly the same experiences.

They had lots of ideas, so we asked AI to generate images of their ideas, including one of young people being tutored. Immediately, they responded: “No, there are too many people in that room; we want one-to-one support” and “That image looks like a school—having tutoring in a school setting can feel embarrassing. Let’s find community spaces instead.”

Using AI helped us understand what the young people wanted far more quickly. And frankly, it was a lot less cringey than asking a facilitator to make a group of artistic novices draw their hopes and dreams.

Would I use AI in a room full of young people without testing it first? Absolutely not. AI has biases, and we need to be mindful of them. I don’t need to ask young people if they think all the people we support should be white, but I do need to prompt AI to reflect Camden’s actual diversity.

Using AI to Reduce the Time Burden on Participatory Grantmakers

Our grantmakers are busy people. They’re studying, working, caring for others, and improving their own lives. They don’t have endless hours to read through stacks of applications.

With caution, we’ve explored AI-powered tools to help make reviewing applications easier. With applicants’ permission, we uploaded proposals to Notebook LM so that grantmakers could listen to audio summaries instead of reading long documents. For some, this was a game-changer—they simply wouldn’t have had time to engage otherwise.

But it wasn’t perfect. We go to great lengths to ensure our panellists aren’t triggered by distressing content in applications. This became harder with Notebook LM, which, in a cheerful American accent, casually discussed gang violence. It felt flippant to a group of young people for whom that topic is anything but flippant. The technology will no doubt improve in handling sensitive content, but for now, it’s something to navigate carefully.

Using AI to Compare Traditional and Participatory Grantmaking

Lastly, we did a bit of an experiment: we used AI to generate grant decisions without telling the young people.

We uploaded all the applications, prompted the AI with the young people's priorities and some grantmaking basics, and asked it to select the top four projects from a shortlist of 13. Meanwhile, the young people carried out our usual participatory decision-making process. When they revealed their choices, two of their top four projects matched AI’s picks.

4 of the Youth Panel engaged in discussions about the grants. Photo credit, Naima Browne.

For the two that aligned, both the AI and the young people valued strong track records, well-evidenced processes, and well-thought-out project designs. But for the two that only the young people selected, their reasoning was rooted in community reputation. Almost every applicant was known to at least one person in the room, and two organisations stood out as doing life-changing work. One person summed it up: “I’ve seen the faces of boys when they come out of there—they’re changed forever by that charity.”

Conversely, for one project that AI liked but the young people didn’t, they loved the idea and recognised its need but felt it was already accessible in Camden—funding it further would be duplication.

This experiment was exciting because it’s helped us better understand why community knowledge leads to different grantmaking decisions. It’s not just about what’s written in an application—it’s about lived experiences and deep, local understanding.

Where We Go From Here

Our ambition at Camden Giving over the next 10 years is to significantly scale the number of people involved in grantmaking. We see this as a crucial part of increasing active citizenship and tackling inequality in Camden.

AI won’t replace community decision-making—far from it. But it can be a tool to make participatory grantmaking more efficient, accessible, and insightful. Used wisely, it can support the shift of power to those who need it most.

Find out about receiving funding from us here.

Find out about our work to shift power in the funding sector here.

Find out how you can help us overcome inequality in Camden here.