AI in UK charities: adoption statistics
Most UK charities now use AI in some way, but governance and skills have not kept pace. In 2026, 79% of charities reported using AI tools, up from 76% a year earlier — yet 51% still had no AI policy, and the most common concern is that AI gets things wrong. This page sets out the sourced picture of how the sector is adopting AI, what it uses AI for, and what worries it.
In 2026, 79% of UK charities reported using AI tools in some way, up from 76% in 2025 — but 51% still had no AI policy in place (Source: Charity Digital Skills Report 2026, Zoe Amar Digital).
Key statistics at a glance
These figures are from the Charity Digital Skills Report 2026, the sector’s main annual survey of charity digital and AI adoption. It is a self-selecting survey of charities, so it reflects those who chose to respond rather than a random sample; see Methodology.
- 79% of charities use AI in some way, up from 76% in 2025.
- 38% use AI actively or strategically in everyday work, up from 25% in 2025.
- 20% use no AI tools intentionally at all.
- 60% use AI for everyday tasks like summarising notes and drafting emails (up from 48%).
- 45% use AI for grant fundraising — rising to 52% among small charities.
- 51% have no AI policy in place (62% of small charities; 21% of large).
- 56% say limited skills and technical expertise is their biggest barrier (up from 43%).
- 1% are reducing staff in response to AI.
How many charities use AI, and how deeply?
Adoption is now the norm, but depth of use is not. In 2026, 79% of charities were using AI in some way, but that splits into 41% still only “exploring”, 34% using it actively in some tasks, and just 4% using it strategically as part of their plans. A further 20% used no AI tools intentionally. So while headline adoption is high, most charities are at an early, informal stage rather than embedding AI across their work.
Size matters. Large charities are well ahead, with 92% exploring, actively or strategically using AI, compared with 74% of small charities. The share using AI actively or strategically rose from 25% in 2025 to 38% in 2026 — real movement, but from a low base.
Chart 1 (for Ivan to build): AI adoption stage (not at all / exploring / active / strategic), 2025 vs 2026. Free to reuse with attribution to CharityIQ and the Charity Digital Skills Report 2026.
AI adoption stage, 2025 and 2026
| Stage | 2025 | 2026 |
|---|---|---|
| Not using AI intentionally | 21% | 20% |
| Exploring | 51% | 41% |
| Active (used in some everyday tasks) | 23% | 34% |
| Strategic (widely used, part of plans) | 2% | 4% |
Figures are exactly as published on the Charity Digital Skills Report’s AI Adoption page (charitydigitalskills.co.uk, checked directly 10 July 2026). Note the 2025 column sums to 97%, not 100% (21+51+23+2); this is a rounding artefact in the source’s own published percentages, not a transcription error on our part or a hidden “don’t know” category we could find on the page — we have reproduced the source’s figures exactly as stated.
What do charities use AI for?
Charities mainly use AI to save time on routine tasks. The most common day-to-day uses in 2026 were everyday tasks like summarising notes and drafting emails (60%), creating documents and reports (57%), research (47%), and generating ideas (43%). Use for grant applications (35%) and analysing spreadsheet data (25%) both grew, and 31% now use AI as a “thinking partner” to work through decisions.
At an organisational level, the top uses were administration and project management (63%) and grant fundraising (45%). Small charities lean on AI for fundraising and social media, where they have least capacity — 52% of small charities use AI for grant fundraising, against 30% of large ones. Use in higher-risk areas stays low: only 9% use AI in service delivery.
Chart 2 (for Ivan to build): top day-to-day and organisational uses of AI. Free to reuse with attribution.
How charities use AI (2026, share of charities)
| Use | Share |
|---|---|
| Everyday tasks (notes, emails) | 60% |
| Creating documents, reports, presentations | 57% |
| Research and information gathering | 47% |
| Generating ideas and creative thinking | 43% |
| Administration and project management (organisational) | 63% |
| Grant fundraising (organisational) | 45% |
| Monitoring and evaluation (organisational) | 31% |
| Service delivery (organisational) | 9% |
The policy and governance gap
Adoption has raced ahead of governance. In 2026, just 18% of charities had a formal AI policy and another 29% were developing one or had agreed good practice — leaving 51% with no policy at all. The gap is widest for small charities, 62% of which had no policy, against 21% of large charities. Alongside this, 44% of charities had taken no action at all to progress AI or manage its risks.
Specific governance steps are rare: only 7% review AI use and risks at board meetings, only 6% have carried out risk assessments for AI use cases such as bias or misinformation, and the share updating their risk register for AI actually fell, from 23% in 2025 to 15% in 2026. This is the gap a simple, adopted AI use policy is designed to close.
Concerns and barriers
Charities are adopting AI thoughtfully, with a clear view of the risks. When asked in their own words, their single most common concern was accuracy — that AI “hallucinates” or makes mistakes, so outputs must be checked, sometimes cancelling out the time saved. Other strong themes were data protection and safeguarding (especially with vulnerable people’s data), the loss of an authentic human voice (“AI slop”), environmental and ethical concerns, cost, and a lack of skills and policy.
On the measured barriers, people — not technology — top the list. The biggest barrier in 2026 was limited skills and technical expertise (56%, up from 43%), followed by lack of training (45%), data protection and security concerns (43%), quality and accuracy concerns (40%), and — notably — lack of trust in AI tools (35%, up sharply from 15%). Concern about colleague resistance also rose, to 34%.
Chart 3 (for Ivan to build): top barriers to AI progress, 2025 vs 2026. Free to reuse with attribution.
Biggest barriers to AI progress (2026)
| Barrier | Share (2026) | Change from 2025 |
|---|---|---|
| Limited skills and technical expertise | 56% | up from 43% |
| Lack of training or support to upskill | 45% | up from 39% |
| Data privacy, GDPR and security | 43% | unchanged |
| Quality/accuracy of AI answers | 40% | up from 33% |
| Lack of trust in AI tools | 35% | up from 15% |
| Colleague resistance | 34% | up from 20% |
| Environmental impact | 31% | up from 26% |
How charities feel about AI
Charities are hopeful but not confident. In 2026, 63% said they were excited about AI’s potential to improve efficiency and impact, but only 41% believed their charity was well set up to adapt to it. Enthusiasm varies by role: 74% of CEOs and leaders were excited about AI, against 47% of trustees, and only 27% of trustees felt their organisation’s culture encouraged experimentation. On the workforce, the evidence so far is reassuring — only 1% of charities were reducing staff in response to AI, and 29% were upskilling staff and volunteers.
Methodology and sources
This is where the figures come from and how far to trust them.
- Source: the Charity Digital Skills Report 2026, produced by Zoe Amar Digital, the sector’s established annual benchmark for charity digital and AI adoption. Figures are quoted from the report’s published findings (updated online July 2026) and attributed; the report is copyright Zoe Amar Communications Ltd, so we summarise and cite rather than reproduce it.
- Survey type: a self-selecting online survey of UK charities, not a random sample. Respondents opt in, so the sample can over-represent charities more engaged with digital — worth bearing in mind, especially for adoption figures. The report compares like-for-like year on year, which makes the trends (2025 to 2026) more reliable than any single number.
- A note on the headline number: the report’s own interim findings (first 240 responses) put AI use at 88%, but the final figure across the full sample was 79%, after a late influx of responses from small charities and those earlier in their digital journey. We use the final 79%, as the report recommends, because it is more representative — a good example of why the sampling detail matters.
- Large vs small: the report consistently finds large charities further ahead than small ones on adoption, policy and governance; we show both where it is informative, since CharityIQ’s audience is small charities.
- [DATA — CharityIQ survey, when available]: when CharityIQ runs its own survey of small charities’ AI use, add those findings here as a clearly-labelled second source, and note where they agree or differ with the Charity Digital Skills Report.
- Sample size for 2026: 807 responses (the survey’s largest ever since it began in 2017), collected between 5 March and 20 April 2026, of which 570 (71%) were from charities with income under £1 million (Source: Charity Digital Skills Report 2026, “Background to the Research” page, charitydigitalskills.co.uk, verified 10 July 2026).
Cite this page
Free to reuse with attribution. Suggested citation:
CharityIQ (2026), “AI in UK charities: adoption, uses and concerns”, summarising findings from the Charity Digital Skills Report 2026 (Zoe Amar Digital). Available at: https://charityiq.co.uk/data/ai-in-uk-charities/ (accessed [date]). Underlying survey findings © Zoe Amar Communications Ltd.
Changelog
- 10 July 2026: first published, using the Charity Digital Skills Report 2026 (figures as published online, updated 9 July 2026). To be refreshed when the next edition publishes, and updated with CharityIQ’s own survey data when available.
The gap this data shows — most charities using AI, half with no policy, and accuracy their top worry — is exactly why CharityIQ is built to be grounded and auditable, with a person checking every output.