AI in Charities is Moving Faster Than Most Policies Can Keep Up
January 28th 2026 | Posted by Emily Formby
AI in charities jumped from 61% adoption to 76% in a single year and Charity Recruit’s latest virtual boardroom set out exactly what that shift means for leaders heading into 2026.
Zoe Amar, founder of Zoe Digital and co-chair of the Charity AI Task Force, walked through the data behind that growth, a case study of a ten-person charity getting real results and the governance gaps most organisations still have not closed.
“AI is one of the biggest challenges facing charity leaders right now, coming at a time when the sector is already under exceptional financial and service pressure.”
Where a Ten-Person Charity Is Already Ahead
Amar’s clearest example came from Speakers Trust, a small charity that sends volunteers into schools to build young people’s presentation skills. Buried in admin, the team combined AI with automation tools to pull daily data on pupil numbers and session feedback into reports that land in staff inboxes automatically.
They then built an interactive AI coach so students can get feedback on a presentation the night before it is due, without replacing the face-to-face volunteer sessions that remain the core of the charity’s work. It is a small organisation proving that scale is not a precondition for meaningful use of the technology.
What a Solid AI Policy Actually Needs to Cover
Only 48% of charities currently have an AI policy, up from 16% the year before and Amar said most of the ones she reviews are still too thin. She pointed to four things every policy needs:
- A clearly defined scope: What the policy actually governs, rather than a vague statement of intent
- Data privacy guidance: Specific rules on what information staff can and cannot put into AI tools
- Clear boundaries on use: Concrete examples of acceptable and unacceptable tasks, not general principles
- A link to inclusion and bias review: A process for checking AI outputs for discrimination before they reach beneficiaries
The Risk Leaders Cannot Ignore
Beyond policy, Amar flagged disintermediation as the sharpest threat facing charities that rely on people finding them directly.
“Beneficiaries are already going straight into ChatGPT, Copilot, Claude or Gemini for advice about their condition and if that information is not accurate or relevant for them, it can cause real harm.”
She also warned against a quieter danger. As AI gets better at tasks that once needed a person, there is a growing temptation to view roles as easy to cut. Amar argued that the sector’s depth of relationship and community knowledge makes most jobs far harder to automate than a spreadsheet exercise suggests and urged leaders to resist decisions driven by cost alone.
Spread the Skills Before You Need Them
On capability, Amar pushed leaders to avoid depending on one or two enthusiastic early adopters. Distributing AI champions across teams and levels, rather than letting expertise sit with a handful of people, protects an organisation against the risk of losing that knowledge overnight if someone moves on.
In Summary
AI in charities has moved from experiment to expectation in the space of a year and the gap between adoption and proper governance is where the real risk sits. Leaders who build clear policies, spread skills widely and treat disintermediation as a strategic issue rather than an IT one will be better placed heading into 2026.
