While the national conversation around AI remains obsessed with efficiency and productivity, a different experience is emerging at the grassroots level. For local communities, the AI transition isn’t an administrative upgrade, and much deeper questions are beginning to emerge around influence, power and agency.
Research suggests marginalised communities face the most significant barriers to accessing the benefits offered by AI. Yet there is an opportunity for communities to harness AI and see more positive outcomes, unique to them. What is the role of these communities in shaping the future of artificial intelligence? How can citizens and civil society groups play a larger role in the design, deployment and governance of AI technologies?
These questions and more were the focus of a recent discussion, hosted by Prospect deputy editor Ellen Halliday and David Knott, CEO, The National Lottery Community Fund—who as part of the Fund’s commitment to enable capability building with communities, has recently announced a
£3 million community AI pot to help them develop and sustain their own power, shaping solutions that are rooted in local strengths and aspirations.
The need for the narrative around AI to be reframed was clear from the outset: “All of the oxygen is being sucked up by talking about the technology solutions and we’re not having discussions about what kind of society we want to live in, what health should look like, what education should look like.”
Others agreed that the potential for communities lies beyond productivity: “Let’s not begin with the idea that people can save half an hour per day doing a task—it’s really about power.”
Harnessing that power would require bridging a gap in capability, confidence, access and influence. As one community representative pointed out, challenges in adoption were not necessarily about resistance to innovation but a rational response to exclusion and a lack of agency: “We see community groups and organisations saying they don’t know where to go to better understand AI, they have concerns about its use and so either end up resisting it completely or adopting it in a way [in which] they feel their values are compromised.”
‘It’s a problem-driven approach, rather than seeing AI as a hammer and everything is a nail’
Overcoming these challenges is not easy, but there are ways communities can be better enabled to become active shapers of AI rather than passive recipients of it. One such opportunity may be found in communities being supported to design and build smaller, more specialised AI models rather than the large, general purpose models that tend to grab the headlines. “Perhaps the things people want to use in their community aren’t the consumer brands,” challenged one participant.
“One action we took in an NHS Trust, using a smaller model, was to work out why people miss appointment slots. We used AI to identify which people were more likely to not turn up and then what was the best time to remind them of the appointment, across two hospitals. We were able to create 700 more appointments each week.”
Another participant agreed: “If it’s possible to decentralise AI models and have different communities create smaller versions of these things we might see the benefits without the huge cost to people and the planet.”
They made the point that by speaking directly to the communities themselves, we may be able to find the things that are of benefit and build on those insights. “We did a survey with black and brown women about their use of AI, and found a large percentage used it to help them articulate their health symptoms in conversations with doctors, ensuring their health conditions were taken more seriously. We wouldn’t have known this without speaking to them.”
A place-based approach could be the answer to unlocking benefits in communities across the country. Conversations with communities should ascertain the assets that they want to build on, the challenges and where they think improvements could be made: “It’s a problem-driven approach, rather than seeing AI as a hammer and everything is a nail.”
It was agreed there is an important role for funders, policymakers, technology companies and others in creating the enabling conditions for this to happen. “Turning those concerns and conversations that are happening within smaller groups at a local level into a different kind of AI system.” It was agreed levers such as procurement, funding criteria, open-source investment, education and experimentation could all help shape a more equitable distribution of power and benefit.
“We are making good progress with community-led AI work,” concluded one leader. “By testing, learning, and listening to communities, we can identify valuable insights, and our role as funders is to help bring those ideas forward and turn them into meaningful outcomes that communities can shape and own.”
