AI arrives as a burden, not a permission slip
In an under-resourced civic organisation, a new AI capability mostly adds work. It only helps if someone takes work away at the same time.
Earlier this year I spent a full day in a room with people who run charities, trying to work out together what AI does to organisations like theirs. I went in carrying the optimistic version of that story, the one told at conferences, where these tools save time and let the smallest organisations do more with less. I left with a picture much closer to what the room had been saying to each other all day.
Most of the civic and voluntary institutions I come across don’t have room to absorb something new. I mean schools, charities, small councils and the committee that runs the local church hall. There’s no budget line for it, nobody whose job includes evaluating tools and no slack in the week to learn one more thing properly. A new AI capability has to be learnt, checked and paid for or found free. It has to work alongside whatever else is limping along, and keep running after the person who set it up has moved on. In most of the organisations I’ve been in, the cost arrives first and doesn’t go away.
Much of what gets written about AI in the voluntary and community sector comes from people who don’t have to run the thing themselves. They aren’t the volunteer keeping the safer-recruitment paperwork current and filing the return a regulator wants back by a fixed date, in the same week as the frontline work. From that distance, a new AI tool looks like time handed back.
From inside an organisation with no paid staff and no money for software, it looks different. A handful of volunteers cover finance, safeguarding and communications between them. To them the same tool is one more vendor pitch to evaluate and one more login to manage. If it goes wrong, it goes wrong in their name, in a system nobody else in the building understands well enough to fix.
The burden that’s already there
Safer recruitment, data protection and regulatory returns fall on people who aren’t paid to do them, on top of whatever the organisation exists to do. An AI tool has to justify its place in that queue against everything else waiting.
Writing about AI adoption in this sector that doesn’t start from there ends up doing something it probably didn’t mean to. It becomes one more voice telling people who are already stretched that they’re falling behind.
The adoption figures that get quoted come from surveys people choose to fill in. An organisation run by volunteers with no spare hours is the least likely to answer, and being missing from the sample doesn’t mean it’s behind.
When it becomes a permission slip
My working rule for digital change is that it has to subtract before it adds. If someone is being asked to learn a new tool, something else needs to come off their plate at the same time, properly, so the total goes down. I’d rather we said no to some things than fall into a default-online posture by accident, and AI gets the same test as anything else with a login screen.
That’s what I mean by a permission slip. The tool lets someone stop doing something, and they end the week with less to carry. I don’t have a clean example of one yet. Most of what I’ve seen is a capability handed over on the assumption that someone downstream will absorb it. A person still has to decide what stops, deliberately, and then stop doing it.
AI can still be worth using in a small, stretched institution, as long as someone has decided what comes off the plate first.
Ben
Photograph by Alexander Grey.