Responsible AI for Charities

Safe use of AI in your charity

Generic AI training tells your staff that AI tools exist. This course tells them which ones your charity has approved, what data is off-limits, and what to do when something goes wrong.

Your staff and volunteers are already using AI tools. Some are approved. Many are not. This half-day practical session builds your charity's approved-use rules, produces four take-away governance artefacts, and gives everyone a repeatable method for working within those rules in their daily charity work.

Duration: Half day, 3.5 hours
Delivery: In-house or online
Group size: Up to 15 participants
Price: £895 ex VAT per session

Book a free discovery call

£895 ex VAT

Half day (3.5 hours)

All staff and volunteers. Particularly valuable for teams already using AI without formal guidance.

Request a booking

The problem this course addresses

"Our staff are using AI tools. I know they are. I've seen grant applications drafted in free ChatGPT accounts, case notes summarised in online tools I've never approved, donor emails written by AI assistants nobody mentioned to me. We don't have an approved list. We don't have a policy. And every time I try to get on top of it, another three volunteers join who've never heard of any of our rules. Generic AI training exists but it teaches people that AI exists. It doesn't tell them what we allow, what they can do with our beneficiary data, how to check an AI output before they act on it, or what to do when one of those outputs is wrong."

That is not an unusual position. 71% of UK employees have used AI tools their organisation has not approved, and 51% do so every week (Microsoft/Censuswide, October 2025). In charities, the problem builds further. High volunteer turnover means trained individuals leave before knowledge transfers to anyone else. Sessional staff and volunteers often work from personal devices outside standard IT controls. The pressure to deliver services with limited time means staff reach for the quickest available tool, approved or not.

Generic training does not close this gap. The DSIT Flexible AI Upskilling Fund evaluation (January 2026) found fewer than half of employees who completed AI training could quality-assure AI outputs afterwards. The problem is not awareness of AI tools. It is the distance between knowing a rule exists and knowing how to apply it when you have a grant deadline in twenty minutes.

Who this course is for

This course is for all staff and volunteers who use AI tools in their work, at any level of experience. That includes fundraising teams, communications staff, programme delivery teams, and volunteer coordinators. If anyone in your charity has used a free AI tool to save time, they are the right participant.

Right for this course	Better served by a different course
Staff and volunteers who use AI tools, approved or not	Trustees needing a governance overview before approving AI use: see AI for Trustees and Charity Leaders
Organisations that want to move from awareness to organisation-specific approved-use rules	Charities whose priority is completing a DPIA and reviewing their data protection position: see Responsible AI and Data Protection
Charities that need documentation to show trustees, funders or regulators

Charities that have not yet covered AI basics will find AI Fluency the right starting point.

What participants will be able to do after this course

Identify which AI tools the charity has approved for each type of work, and explain why other tools fall outside those limits
Apply the correct data classification rules when working with beneficiary, donor, operational or safeguarding-related material in AI tools
Verify AI outputs before publishing, sending or acting on them, checking for factual errors, misleading content, or material that could misrepresent beneficiaries or donors
Follow the escalation procedure when an AI output is wrong, harmful or uncertain, rather than deleting it and moving on
Apply the meaningful human involvement standard required by the Data (Use and Access) Act 2025 when AI is involved in decisions affecting individual beneficiaries
Explain to colleagues why the approved route exists and how to find it quickly in daily work
Use the four take-away artefacts to brief incoming staff and volunteers without depending on anyone's memory of the session

What the course covers

Every session is built from a 30-minute discovery call in which we learn your charity's own tools, approved platforms, data categories, and workforce make-up. No two deliveries are the same.

Approved tool identification

Participants work through which AI tools the charity has approved for each type of task, and understand the reasoning that makes other tools unsuitable. Staff who understand why a limit exists are more likely to apply it than those who are simply told what it is.

Data classification for AI use

The session covers what categories of beneficiary, donor, operational, and safeguarding data can and cannot be entered into AI tools, and why. Participants leave with data classification quick-reference cards built for your specific approved platforms.

Output verification

Practical techniques for checking AI outputs before they reach a beneficiary, funder, or publication. This includes factual accuracy, confident-sounding errors, and content that could misrepresent the people your charity serves. Checking is a teachable skill; most generic training does not teach it.

Escalation and incident reporting

What to do when an AI output is wrong, misleading, or potentially harmful, and how to document it. Deleting an incorrect output and continuing without a record is not a sufficient response.

Shadow AI

Why staff and volunteers reach for unapproved tools, how to reduce the friction that drives that behaviour, and how to build a working culture where the approved route is the easier one. Technology controls cannot cover every personal device or every shortcut taken in good faith. This session addresses the reasons, not just the behaviour.

Workforce churn and institutional memory

Approved-use rules designed to stay useful when individuals leave, rather than depending on anyone's memory of a training day. Skills transfer to people. The four artefacts produced during this course transfer to the organisation.

Meaningful human involvement

The DUAA 2025 requires genuine human review wherever AI is involved in decisions affecting individuals. This means checking AI-drafted fundraising appeals before sending, reviewing AI-summarised case notes before acting on them, and applying scrutiny to AI-generated eligibility assessments for hardship funds. Reviewing an AI summary in thirty seconds is not meaningful human involvement. This session is direct about what the standard requires in everyday charity practice.

Participants also leave with a method guide they use after the course: a repeatable approach to working within your charity's approved-use rules that does not depend on remembering training day.

How the course is delivered

Step 1: Discovery call, 30 minutes

Before the session, we hold a 30-minute call to understand your charity's current AI tools, data categories, approved-use position, volunteer mix, device use, and turnover rate. Every session is built from that starting point.

Step 2: Bespoke session

Delivered in-house at your charity's premises or online via video call, for groups of up to 15. All materials are written for your tools, your data categories, and your approved platforms. Larger teams are split across multiple sessions at the same per-session price.

Step 3: Compliance documentation

Every participant receives documented training completion evidence meeting Information Commissioner's Office (ICO) audit requirements under the UK General Data Protection Regulation (UK GDPR) accountability principle. This includes completion records and assessment evidence that can be presented to regulators, funders, or trustees.

What is included

A 30-minute pre-course discovery call
Bespoke session materials built around your charity's own tools and approved platforms
A method guide participants use after the course, as a repeatable reference
Documented training completion evidence meeting ICO audit requirements
All session materials shared with your charity after delivery
A follow-up call offered to every client after the course
A certificate of completion on request, at no extra charge

Produced during this session

These four artefacts are specific to this course. They are the difference between a training event and a governance output.

Approved tool list: which AI tools your charity has approved for each type of work
Data classification quick-reference cards: what data can go into which tool, for which type of task
Escalation procedure: what to do when an AI output is wrong, misleading or harmful
Incident reporting templates: the documentation process for AI errors and incidents

These artefacts remain with the charity when individuals leave. Individual training transfers skills to people. These stay with the organisation.

How this course aligns with UK and EU regulation

Data (Use and Access) Act 2025

The DUAA 2025 received Royal Assent on 19 June 2025. Where AI is involved in a decision affecting an individual, including beneficiary eligibility assessments, safeguarding referrals, and hardship fund allocations, that person must be informed, must be able to make representations, and must be able to request human intervention. Meaningful human involvement means a person reviews and can challenge the AI-generated decision. It does not mean rubber-stamping an output.

The ICO's Recruitment Rewired report (March 2026) found that most organisations described automated tools as "decision support" when those tools were making the decisions. The same mischaracterisation applies in charity work. Reviewing an AI-summarised case note in thirty seconds before acting on it is not meaningful human involvement. This course addresses what the Act requires in the tasks charity staff carry out every day.

ICO expectations under the UK GDPR accountability principle

The UK GDPR requires organisations to demonstrate accountability for how they process personal data, including processing that involves AI tools. The ICO's Data Protection Audit Framework expects documented training programmes, completion records, and assessment evidence as part of that accountability position. This course produces evidence that meets those expectations. The ICO does not mandate AI training as a standalone legal requirement, but the documentation produced here directly supports a charity's accountability position under UK GDPR.

EU AI Act Article 4: who it binds and who it benchmarks

Article 4 of the EU AI Act (Regulation (EU) 2024/1689) requires providers and deployers of AI systems to ensure their staff have appropriate AI literacy. It has been in force since 2 February 2025.

For charities with EU exposure, this is a binding legal obligation. EU exposure includes serving EU-based beneficiaries, receiving funding from EU sources, or having AI outputs used in EU jurisdictions. For charities operating entirely in the UK with no EU exposure, Article 4 does not apply directly. It sets the benchmark that UK regulatory thinking is moving toward. This course contributes to meeting the Article 4 obligation for EU-exposed charities, and demonstrates the equivalent standard for UK-only organisations.

Evidence and trust

Listed on the UK Register of Learning Providers, Insightful AI Solutions Limited holds UK Provider Reference Number (UKPRN) 10098807.

This course was designed around the regulatory obligations that apply to charities using AI: the DUAA 2025 meaningful human involvement standard, the ICO's audit framework expectations under UK GDPR, and the EU AI Act Article 4 AI literacy obligation. It was not built around what AI tools can do. It was built around how charities must handle their outputs responsibly, and how to evidence that they have done so.

Training is designed and delivered by practitioners. Co-founders Ben Sefton and Kane Lukassen lead our team.

Where this course fits

This course sits fourth in the Responsible AI for Charities programme, after AI Fluency, Responsible AI and Data Protection, and AI Governance and Funder Accountability. No prior course is required. It stands alone and suits charities that want to move from awareness to action and produce a governance output on the day.

Charities that have completed AI Fluency will find this course builds directly on that foundation. For data protection concerns, pair it with Responsible AI and Data Protection. Trustees seeking a governance overview before approving AI use should start with AI for Trustees and Charity Leaders.

Frequently asked questions

What is the difference between this course and AI Fluency?

AI Fluency builds understanding of how AI works and when it is appropriate for charity use. This course builds your charity's specific approved-use rules and ensures staff know how to apply them in daily practice. The two complement each other. AI Fluency is the recommended starting point, but this course stands alone if your team already has basic AI awareness.

Can this session cover paid staff and volunteers together?

Yes. Sessions are designed for mixed groups. High volunteer turnover is addressed directly. The four take-away artefacts produced during the session (approved tool list, data classification cards, escalation procedure, and incident reporting templates) are designed to remain useful as individuals join and leave the charity.

What do we leave with at the end of the session?

Four artefacts built for your charity during the session: an approved tool list, data classification quick-reference cards, an escalation procedure for AI incidents, and incident reporting templates. These are produced on the day, not sent afterwards. They are governance outputs you can share with trustees and present to funders or regulators.

How does this course deal with staff using AI tools we haven't approved?

The session includes a dedicated section on shadow AI. It covers why staff and volunteers reach for unapproved tools and how to build a culture where the approved route is easier to follow. Rules alone do not change behaviour. The session addresses the reasons people reach for unauthorised tools, not just the behaviour itself.

Does this course address our obligations under the Data (Use and Access) Act 2025?

Yes. The DUAA 2025 requires meaningful human involvement where AI is involved in decisions affecting individuals. The course addresses what this means in everyday charity work: reviewing AI-drafted fundraising appeals, acting on AI-summarised case notes, and handling AI-generated assessments. Rubber-stamping outputs does not meet the standard the Act sets.

How many people can attend one session?

Up to 15 participants per session. Larger teams are split across multiple sessions at the same per-session price of £895 ex VAT. A programme discount of 10 to 15% applies when three or more courses are booked together. The discovery call ahead of delivery helps plan the right session structure for your team size, volunteer mix, and working patterns.

How do you build the session around our specific tools and policies?

Before delivery, we hold a 30-minute discovery call to understand your approved tools, data categories, and current AI position. Every element of the session, including the worked examples, approved-use rules, and data classification cards, is built from that conversation. Nothing in the session is generic.

Does this produce evidence we can show the ICO or a funder?

Yes. Every session produces documented training completion evidence meeting ICO audit requirements under the UK GDPR accountability principle. For charities facing funder due diligence on AI governance, this documentation forms part of the compliance record, alongside the four artefacts produced during the session.

Ready to explore what AI can do for your organisation?

Whether you're just getting started or looking to scale, we'll help you find the right path, responsibly.

Book a Free Discovery Call Take the free assessment