AI dominated the digital for charities news & stories in April’s – and not entirely in a good way. Governance, trust, and the gap between what boards think is happening and what staff are actually doing dominated the UK conversation. Meanwhile, Europe’s AI Act negotiations hit a wall, and some useful global fundraising research hardened into consensus. Here is what you need to know.
April Digital News for Charities
The governance problem is now well-documented
The headline story of the month came from Charity Excellence Framework, whose Future Charity Report: AI in the Charity Sector 2026 landed in late April and was covered across Third Sector, Civil Society News and TFN.
The findings are pointed. Three-quarters of charities are using AI in some form, but only 2% are doing so strategically. All three governance indicators assessed – strategy, accountability and training – are rated red. Most striking of all: Charity Commission research from 2024 found that just 3% of trustees believed their charity was using AI at all. That figure is worth sitting with. Staff are using AI tools daily; boards largely do not know it is happening.
The public trust picture is equally sobering. Charity Tracker research cited in the report found that 38% of UK adults consider it unacceptable for charities to use AI to decide who receives support. Only 13% are comfortable with sensitive personal data being processed by AI systems. Charities work in a different trust environment to commercial organisations, and this data shows the public is already forming views on where the line should be.
The report’s core argument is one worth repeating to any board that considers AI governance a job for next quarter: the risk of getting this wrong sits squarely with trustees, who remain legally accountable for AI-related failures in data protection, safeguarding and decision-making, regardless of whether they knew the tools were in use.
AI adoption keeps climbing – but meaningful impact is lagging
The 2026 Charity Digital Skills Report survey closed on 20 April, and the interim findings published by Zoe Amar are striking. 88% of charities are now using AI daily – up from 76% in 2025 and 61% in 2024. Nearly half (46%) say they are using AI actively or strategically, almost double the figure from last year.
There is a more nuanced finding buried in those numbers: the gap between large and small charities is closing. Where large charities once led clearly on AI adoption, the two groups are now converging. Whether that reflects genuine capability-building in smaller organisations, or simply widespread informal tool use that is not yet backed by any strategy, the full report – due in July – should clarify.
The parallel finding from Virtuous and Fundraising.AI – based on 346 nonprofits globally – is telling in context: 92% use AI-enabled tools, but only 7% report major improvements in organisational capability. 81% are using AI individually, without shared workflows. The technology is everywhere; the infrastructure to make it count is not.
Age UK shows what funded AI training actually looks like
One concrete piece of good news: Age UK launched an AI and data academy in April, training 60 employees in partnership with Multiverse. The programme is funded through the charity’s growth and skills levy contributions – a route available to any employer with a payroll above £3 million, and one that sidesteps the perennial problem of finding discretionary budget for digital skills development.
The use cases are practical and fundraising-focused: accelerating live projects, identifying new funding opportunities, streamlining bid writing. This is a model worth examining for any larger charity that already contributes to the levy and has not yet thought about what it could be doing with it.
A warning about AI and your website content
A GOV.UK blog post published on 20 April by the Department for Business and Trade raised an issue that deserves more attention than it has received.
AI overview tools are now indexing every corner of the web, including old, unmaintained pages that organisations have long since forgotten about. In the example cited, an AI tool served up incorrect information about the cost of registering a charity – pulled from an outdated government page that nobody had looked at in years.
The implication for charities is direct. If your website contains guidance pages, resource documents or policy information that has not been reviewed recently, AI tools may be surfacing that content to your beneficiaries, donors and the public as if it were current. A content audit – checking what is live, what is accurate, and what should be redirected or removed – has moved from good housekeeping to something rather more urgent.
The EU AI Act is still unresolved – and the deadline is live
On 28 April, the second political trilogue between the European Parliament, the Council of the EU and the European Commission on the AI Act Omnibus ended without agreement. A third trilogue has been scheduled for 13 May.
The Omnibus proposes deferring the high-risk AI compliance deadline from 2 August 2026 to December 2027. That deferral has not passed. Without formal adoption, the original 2 August 2026 deadline remains in force.
For UK charities, this is mostly a matter of watching the horizon – particularly those with European operations or funding relationships. But any charity using AI in ways that could influence decisions about who receives support, or that processes sensitive personal data through AI systems, should be aware of the direction of travel. Civil society is not sitting passively: Article 19 joined 40 human rights and digital rights organisations in a formal letter to the European Parliament in April, arguing the Omnibus weakens protections for people from high-risk AI systems. That kind of coordinated pushback matters.
Also in April, the European Commission published a one-year review of its AI Continent Action Plan and announced €63.2 million in funding for AI innovation specifically in health and online safety – both relevant for health charities and digital inclusion organisations working in or alongside European partnerships.
What AI fundraising looks like when it actually works
The Chronicle of Philanthropy published a useful piece on 7 April that cut through the prediction cycle to show real deployments.
The College of Charleston has an AI system that contacts alumni by text and email, asks about their personal memories of their time there, records the responses, and then generates personalised appeals that include the details alumni mentioned – the building they loved, the professor who mattered to them. It is not replacing gift officers; it is extending the reach of a team that could never have had those conversations at scale manually.
The Furniture Bank in Toronto used AI coding tools to unify fragmented donation data from multiple systems, saving hours of weekly manual reconciliation. Children’s Miracle Network uses predictive AI to monitor for anomalies – flagging when a regular fundraising team captain who usually signs up in January still has not done so by mid-February, and prompting personalised outreach.
None of this is exotic. All of it requires clean data, clear processes, and someone with ownership of the outcome. The technology is available; the organisational groundwork is the harder part.
On the health side
Two events in April are worth noting for any charity operating in or alongside the health sector.
The King’s Fund held its 2026 Digital Health and AI Conference, focused on practical deployment at scale rather than pilots – a useful corrective to a sector that has accumulated a great deal of evidence that AI can work in clinical settings, and rather less experience of making it work routinely across an NHS trust. Health charities are specifically included in the audience, at concessionary pricing.
On 29 April, the UK’s first Health + AI Tech Show took place in London, supported by equality charity Diversity UK. The three-track programme covered clinical AI and diagnostics, genomics and drug discovery, and AI in health operations. Registration was free for NHS and public sector professionals, and the framing was explicitly about what is already deployable – not what might eventually be possible.
The thread running through all of this
April’s news has a consistent message: AI use in charities is growing fast but governance, training and board-level oversight are not keeping pace. The public has conditional tolerance for AI in charity work, but that tolerance drops sharply when AI touches decisions about who gets help, or handles sensitive data without visible human accountability.
The charities that will convert this moment to their advantage are not necessarily the ones moving fastest. They are the ones building the internal infrastructure – policies, training, board ownership, and clean data – that makes responsible AI use legible to staff, to donors, and to the people they serve.