# AI Transparency For Councils: Telling Residents How You Use AI

URL: https://www.starien.com/blog/ai-transparency-for-councils-telling-residents-how-you-use-ai
Published: 2026-10-30
Author: Geoffrey Moffett
Topics: transparency, responsible AI, councils, AI, public trust

> How councils can be open about their use of AI: a public AI use statement, labelling chatbots and images, what to publish about significant tools and how to answer questions honestly.

## Quick answer

AI transparency for councils means being open about where and how AI is used, especially where it affects residents. Practical steps include publishing a plain-English AI use statement, telling people when they're talking to a chatbot, labelling AI-generated images, publishing information about significant AI tools and answering questions honestly.

## Key takeaways

- Openness builds trust; secrecy invites suspicion when AI use comes to light.
- A short public AI use statement covers most questions.
- Always tell people when they're interacting with AI or seeing AI-generated images.
- Significant tools that affect residents deserve more detailed public information.

Picture a local journalist asking: "Is the council using AI to write its press releases? To answer complaints? To decide things about people?" If the honest answer is "we're not sure", that's a bad story waiting to happen. AI transparency for councils is about having a clear, confident answer ready, and publishing it before anyone asks.

## Why AI Transparency For Councils Matters

- **Trust.** Residents are wary of AI in public services. Openness reassures; discovery after the fact damages trust.
- **Accountability.** Being clear about how AI is used helps show that people remain responsible.
- **Members' confidence.** Elected members can answer constituents' questions.
- **Good practice.** Public sector guidance, including the UK Government's AI playbook for the public sector and the ICO's guidance on AI and data protection, emphasises transparency.

## Three Levels Of Transparency

### Level 1: A Public AI Use Statement

A short, plain-English web page. Something like:

> **How we use AI**
>
> We use AI tools to help staff with tasks like drafting documents, summarising published reports, making information easier to read and planning communications. A member of staff checks and approves everything before it's published or sent.
>
> **What we don't use AI for:** making decisions about individuals' services, [other commitments your council is confident it can keep].
>
> **Your data:** we only use approved tools, and we don't enter personal information unless a tool has been assessed and approved for that use. [Link to privacy notice.]
>
> **When you're talking to AI:** we'll always tell you.
>
> **Questions?** Contact [team].
>
> Last reviewed: [date].

Only commit to things your council actually does and can keep doing. A statement that turns out to be untrue is worse than none.

### Level 2: Labelling At The Point Of Use

- **Chatbots:** say clearly at the start that it's an AI assistant, what it can help with and how to reach a person.
- **AI-generated images or video:** label them, and never present them as real photos of real places, people or events.
- **Synthetic voices:** say so if a voice is AI-generated.
- **AI-assisted translation:** say if a translation was produced with AI and checked by a person, where your council's language policies require it.

### Level 3: Detail On Significant Tools

For AI tools that significantly affect residents, such as a chatbot handling service queries, or any tool supporting decisions, consider publishing more detail:

- what the tool does and why;
- what data it uses;
- how accuracy and fairness are checked;
- who is accountable;
- how to challenge or ask questions.

The UK Government's algorithmic transparency recording standard offers one model for structuring this.

## Routine Drafting: Do You Need To Label It?

Generally, no. If an officer uses AI to help draft a press release, then edits, checks and approves it, it's the council's release. Labelling every AI-assisted document creates noise and can imply the content is less reliable. The key commitment is that a person stands behind it, which your AI use statement explains.

But be honest if asked. "Yes, our team uses AI tools to help with first drafts, and every release is checked and approved by an officer" is a perfectly good answer.

## Preparing For Questions

Have a few approved lines ready for media, members and residents:

- Do you use AI? Where?
- Does AI make decisions about people?
- Is residents' data put into AI?
- Who checks AI's work?
- What happens if it gets something wrong?

AI can help you draft these lines from your policy. A senior officer approves them.

## Making Transparency Part Of The Process

- Add a "transparency" step to your AI risk assessment: what will we tell people?
- Review the AI use statement when you introduce significant new tools.
- Include AI in your annual reporting where appropriate.
- Train frontline staff to answer simple questions about AI use.

## A Made-Up Example: The Journalist's Question

Imagine a local reporter emails the press office: "Readers are asking whether the council uses AI. Does it, and for what?"

**Without preparation,** the press office scrambles to find out who's using what, chases IT and information governance, and replies two days later with something vague. The story becomes "council unclear on AI use".

**With preparation,** the press officer sends a link to the council's published AI use statement within the hour, plus an approved line: "Like many organisations, we use AI tools to help staff with tasks such as drafting and summarising. Everything is checked and approved by a member of staff before it's published. We don't use AI to make decisions about individuals, and we tell residents when they're talking to an AI assistant." The story, if there is one, is "council sets out how it uses AI".

## Keeping The Statement True

A transparency statement is a promise. To keep it true:

- give it an owner, usually within the AI governance group;
- check it whenever a new AI tool or significant use is approved;
- review it at least annually;
- make sure frontline staff, members and the press office all know where it is.

If something changes, update the statement before residents notice the gap.

## Pitfalls To Avoid

- **Over-promising.** "We never use AI" is rarely true once software includes AI features.
- **Jargon.** Write the statement for residents, not for IT.
- **Hiding chatbots.** People feel misled when they discover a "person" was a bot.
- **Fake photos.** An AI image passed off as real can undermine trust in everything else you publish.

For the principles behind this, see [what responsible AI means for council decision-makers](/blog/what-is-responsible-ai-a-framework-for-council-decision-makers/).

## How Starien Can Help Train Council Teams

Our [Public Relations](/public-relations/) background means we think about how AI use will look from the outside, and our [AI Coaching](/ai-coaching/) helps comms and policy teams write clear statements, labels and approved lines. We've worked with public sector teams and know trust takes years to build. [Get in touch](/contact/) to talk about training.

## Frequently asked questions

### Do councils need to label every AI-assisted document?

Not usually. Routine drafting that a person has edited and approved is the council's own work. Focus labelling on chatbots, AI-generated images or video, and significant tools that affect residents.

### What should a council AI use statement include?

Why the council uses AI, the main types of use, what it doesn't use AI for, how data is protected, how people stay accountable, and how residents can ask questions.

### Is there a standard for publishing information about public sector algorithms?

The UK Government has developed an algorithmic transparency recording standard used across parts of the public sector. Councils may find it a useful model, alongside any NI public sector guidance.
