A recap on AI skills
Earlier in our report (see the section on ‘Core digital skills and capacity’), we discovered:
- Skills to use AI tools in everyday work are fairly low: Over a third (35%) say they are poor at this, while 29% say they don’t do this. This means that almost two-thirds of charities (64%) either struggle with or are not using these tools.
- Charities are struggling to keep up to date with emerging tech trends: Reinforcing this trend, half (50%) of charities are either finding this a challenge or are not across these trends. More than 4 in 10 (41%) of charities are poor at this and 9% don’t do this.
However, despite their limited skills and capacity, our extended questions on AI have uncovered increasing adoption and use of AI tools, particularly among large charities.
Use of AI
What stage are charities at with AI?
76% of charities are using AI, compared to 61% who were using it in 2024.
For the first time, we asked charities which stage of AI development they have reached, so we can begin to track how AI maturity is developing across the sector.
- 76% of charities are using AI. This comprises 2% who are either using it at strategic level or planning to scale this, 51% who are exploring tool usage (45% were using it informally last year) and 23% who are piloting and using AI.
- This is a significant increase on the 61% of charities who were using AI in our 2024 report.
- 89% of large charities are using AI, compared to 72% of small charities.
- 91% of charities with a digital strategy in place (at the advanced or advancing stages with digital) are using AI tools.
- Surprisingly, 62% of charities at the earliest curious stage with digital (with limited digital skills and capacity) are using AI tools in their organisation. However, the majority are still exploring AI tools (48%) rather than proactively implementing these in their organisation (13%).
- Infrastructure organisations report higher use of AI, with 36% at the active use stage, compared to 23% in the overall sample.
How are charities using AI day to day?
We’ve seen significant growth since 2024, with half of charities (50%) using AI for documents and reports, as well as for administrative tasks (48%) and idea generation (35%).
The top 5 ways charities are using AI day to day are:
- 50% are using AI for documents and reports (e.g. drafting or structure content).
- 48% are using AI for administrative tasks, up significantly from 32% last year.
- 36% are using AI for developing online content, similar to 2024.
- 35% are using AI for idea generation, up from 27% last year, perhaps indicating growing confidence among charities.
- 29% are using it for research and information gathering, up from 24% last year.
Charities are rarely using AI for tasks involving data analysis, with only 15% using it for qualitative data analysis, 13% for numerical data and 4% for predictive analytics. Later on in this report, we also highlight how charities’ top request for data support is now using AI tools to analyse our data (45%). This is a growing area of interest.
Large and small charities are using AI tools in a similar way in day-to-day work. However, large charities are using AI tools more extensively. For example, 64% of large charities are using AI tools in administrative tasks, compared to 43% of small charities.
How are charities using AI organisationally?
Over a third (36%) of charities are using AI for grant fundraising along with communications and fundraising.
The top 3 ways charities are using AI at organisational level are:
- Administration and project management (48%, up from 32% last year).
- Grant fundraising (over a third 36%, this rises to 43% of small charities, vs only 27% of large charities).
- Communications and fundraising (34%, almost doubled from 19% last year – this rises to 38% of small charities, vs 28% of large charities).
Overall, charities are clearly exploring applications for AI, yet, in reality, few are using it across their organisation. Only small proportions of charities are using AI for:
- Impact and evaluation (e.g. reporting) (17%)
- Service delivery (7%, down from 12% last year)
- Strategic planning (e.g. scenario modelling, predictive data analysis, resource planning) (9%)
- Recruitment and HR (5%). This helpful article from Charity Digital outlines the potential risks of using AI in charity recruitment.
Interestingly, those charities at an early stage with digital (with limited digital skills) are more likely to be using AI tools for:
- Grant fundraising (50% of those at the earliest curious stage, vs 20% of those at the most advanced stage)
- Communications and fundraising (38% of those at the earliest curious stage, vs 29% of those at the most advanced stage).
Moving forward with AI
How are charities feeling about AI?
Almost three-quarters (74%) of charities now agree that AI developments are relevant to them, while 59% are worried about the implications of using AI.
We know from our 2024 report that many charities feel worried, excited about and unprepared for AI (also see CAST’s 2025 AI research and the Joseph Rowntree Foundation 2024 report on Grassroots and nonprofit perspectives on generative AI. This year, we asked further questions to track and understand the mindset with which charities approach AI technologies and how prepared they feel about the changes they bring.
- 74% either strongly agreed or agreed that AI developments are relevant to them – up from 65% last year.
- Almost two-thirds (62%) are excited about the potential for these tools to increase their impact.
- More than half (59%) are worried about the implications of using AI (e.g. for service quality, data privacy). This rises to 74% of large charities, who are further ahead with AI use (vs 53% of small charities). This indicates that worries around AI increase with the use and uptake of AI.
- It’s encouraging that 42% are responding to the opportunities and challenges surrounding AI, given that only 22% felt ready to respond in 2024.
- Only 33% don’t know how to get started with AI. There is a strong link to digital skills and capacity generally here, with 47% of those at the earliest curious stage of digital saying they don’t know how to get started with AI.
How are charities managing the risks in AI adoption so that they can adopt AI responsibly?
Three times as many charities as last year are developing an AI policy and the number of charities reviewing their data protection and security has almost doubled.
This was a new question so that we could track how charities are managing risk in AI.
The top actions charities are taking are:
- Almost half (48%) are developing an AI policy, a number that has tripled since 2024 when only 16% were doing so. This rises to 68% of large charities, compared to 37% of small charities.
- 39% are ensuring that AI aligns with their organisational values and policies, rising to 51% of large charities.
- 36% are avoiding use of AI in potentially harmful areas.
- 36% are reviewing data protection and security (only 15% were doing this last year). This rises to 48% of large charities.
However, it is worrying that just 2 out of 10 charities are taking the following actions, which in our view should be standard, ongoing practice: updating their risk registers (23%), reviewing their governance to give trustees better oversight of AI (22%) and undertaking regular assessments of AI risks (e.g. bias, privacy and misinformation) (20%). These actions are essential to good governance and the responsible adoption of AI.
What actions are charities undertaking to move forward with AI?
Worryingly, almost 4 out of 10 charities (37%) have not undertaken any actions to move forward with AI. This was the most popular option. There is a risk that many will continue to use AI in day-to-day work tasks (and stay stuck in the exploring stage), without looking at this strategically or actively weighing up the risks and benefits of implementing AI tools across their organisation.
The top actions undertaken by charities are as follows. It is notable that even the most widely selected options are only being undertaken by 2 out of 10 charities at most.
- Upskilling staff and volunteers (22%).
- Developing an AI strategy (22%).
- Reviewing ethical and environmental concerns (21%).
- Exploring how AI might reshape roles and responsibilities (15%).
- Establishing an internal AI working group (14%).
It is striking that only 9% are partnering with other sectors and only 4% are hiring staff with AI skills and expertise. While only 13% say they are seeking funding to move forward with AI, this rises to 35% of black led charities.
Large charities are much more active in taking steps to move forward with AI. They are:
- Developing an AI strategy (35%)
- Upskilling staff and volunteers for AI adoption (32%)
- Establishing an internal AI working group (30%)
- Reviewing ethical and environmental concerns (to move forward with AI) (24%)
- Planning for AI developments (e.g. AI agents automating tasks) (19%).
What sector-wide efforts would charities like to see to ensure successful AI adoption?
More than 60% of charities want more guidance on responsible AI, free resources for the sector and training for leaders and trustees.
We asked this new question to understand what charities want to happen across the sector so that support, funding and resources can be co-ordinated.
The most popular needs among charities are as follows:
- 69% want more guidance on what responsible AI adoption looks like.
- 64% want more free resources for charities, indicating that they need guidance that is bespoke and relevant to the sector.
- 60% want training for leaders and trustees.
- More than half (56%) want charities to co-ordinate to share skills, use cases and pitfalls.
Charities also want funders to take the following actions:
- Half (49%) want funders to provide support for skills development and capacity.
- 45% want funders to provide guidance on their AI policies.
- More than a third (35%) want co-ordinated investment from funders.
It is also notable that more than 4 out of 10 (43%) want guidance from regulators and just under a third (31%) want to see more cross sector partnerships.
- Small charities are more likely to say they need direct funder support, with half (53%) saying they need funder support initiatives to develop their skills and capacity in order to move forward with AI.
- In comparison, large charities are looking for sector-wide support. The majority (66%) say there needs to be more co-ordination between charities to share skills, use cases and pitfalls, followed by more guidance from regulators (49%) and a sector-wide plan for adoption and support (45%).
Leadership skills and AI
Just over a third of charities (36%) say that their CEO has poor AI skills, knowledge and confidence, with 4 out of 10 (44%) stating that their board is also poor in this area.
Ever since this report began in 2017, we have seen how crucial the support of leaders and trustees is to digital transformation. Continuing this theme, we wanted to find out how charities rate the AI skills, knowledge and confidence of their CEO and board. The results are worrying.
CEO skills
- 36% say their CEOs have poor AI skills. 39% say that their CEO is fair at AI, and less than 1 in 10 (8%) see their CEO as excellent.
- Only 3% of charities say that their boards are excellent at AI, which is unsurprising given the significant digital skills gap we have seen among trustees ever since this report began in 2017.
- Both large and small charities give their CEO and board AI skills similar ratings.
- Later on in this report, we see that in order to progress with digital, 41% of charities want their CEOs to keep up to date with emerging trends, tech and AI tools, while a third (34%) want their CEO to prioritise understanding risks and opportunities for emerging tech.
Board skills
- 44% say that their boards are poor at AI, with just over a quarter (26%) saying that they are fair.
How charities are learning about AI
More than half of charities are learning about AI by testing out tools in their job and a similar number are attending events in order to grow their skills.
The most popular ways in which charities are learning about AI are as follows. In particular, the top 3 methods have seen strong growth since 2024.
- Almost 6 out of 10 (59%) are learning on the job, i.e. experimenting and testing AI tools at work. This has gone up significantly from 35% last year. This rises to 74% of those working in large charities, compared to 53% of small.
- More than half (56%) are attending events, workshops and webinars – up from 37% last year. This rises to 78% of those working in large charities.
- More than a third (36%) are learning from their peers or other nonprofits, which has grown from 20% in 2024. This rises to 50% of those working in large charities.
- 29% are actively following AI developments – this is similar to last year, rising to 46% of large charities.
- 29% are attending training – up from 12% last year.
Overall, we can see that greater proportions of large charities are learning about AI. This is likely to be linked to their motivation to use AI, capacity available for learning and their digital skills generally.
It’s surprising that only 6% are engaging clients, users or communities about their use of AI. This chimes with the finding from a previous question in this year’s report about the next steps charities are taking in AI: that only 16% of charities are being transparent with external stakeholders (e.g. clients, funders and donors) about AI use. Charities would benefit from increased insights about how their clients are using AI (25% identify this as a support need in a later question).
Which barriers are preventing charities from adopting AI further or using it more?
Concerns about data privacy, a lack of training and worries about factual accuracy, bias and discrimination are the greatest hurdles charities face to AI progress.
The most significant barriers to charities’ AI progress are:
- Data privacy, GDPR and security concerns (43%), up from 31% last year. This rises to 56% of large charities, who say this is the biggest barrier they face.
- Limited digital skills and expertise (43%). It’s positive that this has fallen slightly from 50% last year. This rises to 47% of small charities.
- A lack of training or support to upskill (39%). This rises to 45% of small charities.
- Concerns about factual accuracy blocks a third of charities (33%), rising to 42% of large charities.
- The potential for bias and discrimination (32%). For large charities, this is the second biggest barrier they face, preventing 46% from moving forward with AI. This is also a key barrier for LGBTQIA+ led charities (49%, vs 32% overall).
Only a small proportion of charities (2 out of 10) are being held back at the moment by:
- The costs of resourcing AI tools (29%)
- Ethical and human rights concerns (29%), similar to last year
- 28% don’t know how to get started
- Negativity towards AI among colleagues (20%)
- 26% have concerns about the environmental impact of AI tools
- 25% say that their data and data maturity need attention first, similar to 2024.
Small charities struggle to move forward with AI primarily due to a lack of skills and training, as well as data privacy concerns. Their biggest barriers are:
- Limited digital skills and technical expertise (47%).
- Lack of training or support to upskill ourselves (45%).
- Data privacy, GDPR and security concerns (38%).
Large charities struggle to move forward with AI due to data concerns, as well as worries about the accuracy and bias built into AI tools. Their biggest barriers are:
- Data privacy, GDPR and security concerns (56%).
- The potential for bias and discrimination (46%).
- Issues with factual accuracy in AI-generated answers (42%).
Interestingly, large charities are further ahead in their use of AI in comparison to small charities and yet more large charities are facing barriers to moving forward with AI. Specifically, higher proportions of large charities (compared to small) are struggling to move forward because of data privacy (56%, vs 38%), accuracy (42%, vs 30%), bias (46%, vs 26%) and environmental concerns surrounding AI tools (35%, vs 23%). They also face a wider range of internal barriers, such as negativity among colleagues (32%, vs 15% of small) and data maturity needing attention first (32%, vs 21% of small).
Which aspects of AI would charities most like external support or training for?
More than half of charities want to seek support to grow their understanding of AI and how charities are using it and to explore practical use cases in their work and in their services.
The areas of AI that over half of charities would most like external support for are:
- General understanding of AI and how charities are using this (58%).
- Practical uses of AI in everyday work (58%).
- Exploring potential use cases for AI in services (51%).
- Assessing AI risks (bias, safeguarding) and using AI tools responsibly (51%), a big increase from 40% last year.
Interests in developing AI skills are clearly wide ranging. In addition to the above, 45% of charities want support with using AI tools in their operations and 43% with developing AI policies and governance.
Just under a third of charities (32%) want to learn from their peers to share challenges and ideas with other nonprofits, up from 22% last year.
Small charities most want to develop a general understanding of AI and how charities are using it (63%) and practical uses of AI in their everyday work (58%).
Large charities are also interested in practical uses of AI in their everyday work (61%), but more are looking for support with assessing AI risks and using AI tools responsibly (55%), as well as exploring potential use cases for AI in their services (55%). However, all other support wanted is relatively similar to that of small charities.
