Responsible AI Use Policy

Last updated: 12 July 2026

Neuro Unity uses artificial intelligence in some of our work. We think it can be a genuinely useful tool, and for many neurodivergent people it is an accessibility tool in its own right. We also think that AI can go badly wrong for neurodivergent people if it is used carelessly, because the data behind it often reflects a world that was not built with us in mind.

This policy sets out how we use AI, how we will not use it, and what you can expect from us. We would rather be open about it than quietly use these tools behind the scenes.

1. Scope

This policy applies to everyone working for or with Neuro Unity, including our team, our advisory council, our volunteers, our champions and ambassadors, and any contractor or partner delivering work in our name. It covers all AI tools, including generative AI tools that produce text, images, audio or video, and any analytical or automated tools we use in our services.

2. Our Principles

Human accountability. A named human being is always accountable for anything we publish or deliver. AI does not sign off our work. We do.

Transparency. We tell people when AI has played a meaningful part in something we produce or in a service we deliver. We do not pass off AI generated material as lived experience.

Lived experience comes first. Nothing about us without us. AI does not speak for neurodivergent people, and it is not a substitute for the voices, testimony and expertise of our community.

Fairness and bias. AI systems can reproduce and amplify bias against neurodivergent and disabled people, including in recruitment, monitoring and assessment tools. We treat that risk as real. We check AI outputs for stereotyping, deficit based language and assumptions about how people should communicate, behave or work.

Accessibility. Where AI helps make our work more accessible, for example through plain English summaries, alternative text, captions, transcripts or reading support, we will use it for that purpose and we will keep improving how we do it.

Privacy and data protection. We protect the personal data of the people who trust us with it. Our use of AI never overrides that responsibility.

Accuracy. AI tools produce confident sounding text that can be wrong. Everything factual we publish is checked by a person before it goes out.

3. How We Use AI

We may use AI to:

  • help draft, edit, structure or summarise content, which is then reviewed and approved by a member of our team
  • produce plain English versions, summaries and alternative formats of our resources
  • generate alternative text, captions and transcripts to improve accessibility
  • support research and signposting, where the results are checked against primary sources before we rely on them
  • support internal administration, for example scheduling, note taking and drafting routine correspondence
  • support accessibility and usability testing as one input among several, alongside human review and, wherever possible, testing with neurodivergent people

4. How We Will Not Use AI

We will not:

  • use AI to make or automate a decision about an individual person, for example about their support, their suitability, their diagnosis or their access to our services, because those decisions are made by people
  • offer AI as a substitute for medical, clinical, diagnostic or therapeutic advice, or use it to assess or diagnose anyone
  • enter special category or sensitive personal data, including information about someone’s health, diagnosis or neurodivergence, into a public or consumer AI tool
  • use AI to generate a synthetic voice, image or likeness of a real person without their explicit written consent
  • publish AI generated quotes, testimonials or case studies and present them as real, and every testimonial we publish is a real person’s words, signed off by them
  • use AI to monitor, score or surveil individuals
  • use AI in a way that replaces the paid or voluntary contribution of a neurodivergent person whose lived experience is the point of the work

5. When We Tell You

We will label content where AI has been used to produce a substantial part of what you are reading, watching or listening to. We do not label routine use such as spellchecking, grammar suggestions or scheduling, in the same way we would not label the use of a word processor.

6. Data Protection And AI

Before we use a new AI tool in our work we consider what data it will process, where that data is stored, whether the supplier uses inputs to train their models, and whether the use is compatible with UK GDPR and our privacy policy. Where the use is likely to be high risk we will carry out a data protection impact assessment.

Unity Group Solutions Limited, of which Neuro Unity is an initiative, is registered with the Information Commissioner’s Office under registration number ZB700589.

The AI tools we currently use are Claude, ChatGPT, Google Gemini, Microsoft Copilot and Canva. This list will change over time, and we will update it when it does. We are naming them because we would rather you knew.

7. What We Ask Of Partners And Suppliers

If you deliver work in our name, or you use AI as part of a service you provide to us, we expect you to follow this policy or an equivalent standard of your own, and to tell us where AI plays a meaningful role in what you deliver.

8. Raising A Concern

If you think we have used AI in a way that is unfair, inaccurate, exclusionary or simply not in keeping with our values, please tell us at hello@neurounity.org.uk. We will look into it, we will tell you what we find, and if we got it wrong we will say so and put it right.

9. Review

AI is moving quickly, and this policy will not stay right for long if we leave it alone. We will review it at least once a year, and sooner if the law, the technology or our own practice changes.