The Weekly Scan for week of 6 July ’26
This week’s evidence points the same direction from several angles at once. UK adoption data shows charities using AI more than ever while trusting it less than they did a year ago, a UN gathering in Geneva put governments from every country round the same table to talk about AI safeguards, and European regulators kept adjusting the runway for the AI Act rather than the destination. Underneath all of it sits the same question charity boards are being asked to answer for themselves, whether their governance is keeping pace with their adoption.
CASE STUDY
North America
Chronicle of Philanthropy · North America · 8 July 2026
Nonprofits are adopting AI faster than they can explain it to donors
A new Blackbaud Institute survey of more than a thousand donors and over a thousand nonprofit staff finds a wide gap between what supporters expect on AI and what organisations believe they are delivering. Most donors said clear disclosure when AI is used in donor communications matters to them, and most also wanted assurance that any sensitive data processed by AI is properly protected, yet only a minority of staff believed their own organisation was meeting either expectation. The piece also profiles Fort Collins Habitat for Humanity, where an AI agent is taking on a growing role in donor outreach ahead of a planned expansion into new donor acquisition.
Why it caught our eye: The gap here is not really about the technology, it is about the story an organisation tells its supporters while adopting it. A donor base that is more concerned about disclosure and data protection than staff realise is exactly the kind of blind spot proportionate governance is meant to catch early.
SECTOR RESEARCH
United Kingdom
Civil Society · UK · 9 July 2026
UK charities trust AI less the more they use it
The newly published Charity Digital Skills Report 2026 finds that 79% of UK charities now routinely use AI, up from 76% a year earlier, but the share reporting a lack of trust in AI tools has more than doubled to 35% from 15%. Trust concerns run highest among larger charities and those supporting marginalised communities, and the report notes that a third of boards and a quarter of chief executives have poor AI skills, with 46% of charities lacking a trustee with digital expertise.
Why it caught our eye: Adoption climbing while trust falls is precisely the pattern proportionate governance exists to correct. Left unaddressed, that gap tends to surface later as reputational risk with donors and beneficiaries rather than as a technical problem anyone saw coming.
PRACTICAL GUIDANCE
United Kingdom
Directory of Social Change · UK · 9 July 2026
Charity AI governance moves from IT policy to board agenda
The Directory of Social Change’s second annual Charity AI Conference this week brought together charity leaders to discuss how AI oversight now sits alongside the 2025 Charity Governance Code’s expectation that boards actively engage with technology risk. Sessions ranged from a charity’s own account of adopting AI successfully through to a dedicated session on what trustees, funders and the public now expect of a board when AI enters the organisation.
Why it caught our eye: As one speaker put it, AI oversight “stops being a technical issue and becomes a leadership responsibility.” That is the shift we keep returning to with clients, governance conversations that start with the board rather than the tool tend to age far better.
GLOBAL GOVERNANCE
Europe
UN News · Europe · 7 July 2026
The UN’s first global AI governance dialogue names children’s safety as a priority
The United Nations convened its first Global Dialogue on AI Governance in Geneva this week, bringing all 193 member states together for the first time specifically to discuss international AI safeguards. Secretary-General António Guterres called for common safety standards, human rights red lines and support for developing countries’ AI capacity, and urged nations to adopt a dedicated pledge on AI and children’s safety, while the General Assembly President highlighted the disproportionate way deepfake abuse currently targets women and girls.
Why it caught our eye: For any charity working with children or other vulnerable groups, a global push toward human-rights-grounded AI governance, with children’s safety named explicitly, is worth tracking well before it becomes binding law. Proportionate governance at charity level increasingly sits inside a much larger international conversation about what AI owes to the people it affects.
REGULATION & POLICY
Europe
Fladgate · Europe · 6 July 2026
The EU AI Act’s toughest deadlines move, its direction does not
The Council of the EU has given final approval to a package simplifying the AI Act’s timeline, pushing the toughest high-risk system obligations back to December 2027 while the general-purpose AI regime’s information and enforcement powers still take effect from August 2026. The Commission has also published a final Code of Practice on labelling AI-generated content, including standard EU icons that deployers can use to mark it.
Why it caught our eye: A delayed deadline is not a pause in governance expectations, it simply changes the runway available to prepare. Charities and suppliers working across UK and EU jurisdictions are better served by building toward the Act’s substance now than by waiting for each date to arrive.
REGULATION & POLICY
United Kingdom
Bird & Bird · UK · 5 May 2026
ICO guidance offers charities a clearer line on automated decision-making
The ICO has opened consultation on draft guidance that sets out more clearly how the amended UK GDPR rules on automated decision-making and profiling apply in practice, including what counts as a “decision” and when human oversight needs to be built in. Alongside it sits a report on recruitment, drawn from engagement with over 30 employers, which found that automated decision-making is often woven more deeply into everyday hiring, including shortlisting and scoring, than organisations had assumed. The draft guidance will help inform a forthcoming statutory Code of Practice on AI and automated decision-making.
Why it caught our eye: Many charities use similar tools in their own recruitment, and in grant assessment or service triage too, so this is useful reading regardless of whether AI feels like a live issue for HR yet. Getting a clear picture of where automated decision-making already sits, and where a human needs to stay firmly in the loop, is easier to do now, on your own terms, than later against a checklist.
A Curia AI perspective on the Charity Digital Skills Report 2026
The Charity Digital Skills Report 2026, published this week, is once again one of the most valuable pieces of research the sector produces, and it is worth thanking Zoe Amar, Nissa Ramsey and everyone behind it, along with this year’s partners, for continuing to invest in something that gives charities, funders and suppliers a genuinely evidence-based view of where things stand. The headline will understandably draw attention, adoption up to 79% from 76% last year, with more than a third of charities now using AI actively or strategically rather than simply exploring what it might do. The more interesting story sits beneath those numbers.
Conversations with charity leaders have shifted over the last few months. Very few are still asking whether they should adopt AI, most crossed that bridge a while ago. The question now is how they build the confidence to use it well, how they decide where the boundaries should sit, and how they make sure trust grows alongside capability rather than trailing behind it. This year’s report reads like a sector that has reached exactly that point. The barriers charities now identify are less about whether AI is available to them and more about whether the organisation has the skills, leadership, training and trust needed to use it well, and only one in ten charities say they are planning for what comes next.
That is a capability challenge as much as a technology one, which is why so much of Curia AI’s own thinking has settled around three words.
Principles set the shared boundaries within which AI should operate. People make sure those principles are lived by every trustee, leader and member of staff, not just written into a policy document. Capability brings together the skills, governance, data readiness and ways of working that let an organisation scale AI with confidence. Read against this year’s report, the task ahead does not look like persuading more charities to adopt AI. It looks like helping the organisations that already have, build the principles, develop the people and strengthen the capability that let them scale responsibly, in a way that serves their mission and earns the trust of the people they exist to support.
