How Arts Organisations Can Write a Simple AI Policy

An AI policy does not need to be a 40-page document full of technical language.

For many arts organisations and small charities, the first version can be simple, practical, and easy for staff to use.

The point is not to look impressive. The point is to help people make better decisions.

Why a simple AI policy helps

Your team may already be using AI.

They might be using it to draft emails, summarise notes, brainstorm marketing copy, plan workshops, review funding applications, or tidy up reports.

If there are no shared rules, everyone has to guess what is safe. That is where risk appears.

A simple policy helps answer:

• what can we use AI for?
• what must we not put into AI tools?
• who checks the output?
• how do we protect personal data?
• how do we avoid inaccurate claims?
• what needs a human decision?
• when do we tell people AI has been used?

Start with use cases

Do not start by copying a generic AI policy from the internet.

Start by listing the ways your organisation might actually use AI:

• funding applications
• reporting and evaluation
• board papers
• project planning
• comms and marketing
• education resources
• access documents
• meeting notes
• internal process documentation

Then decide what is low risk, what needs caution, and what is not allowed.

Set clear red lines

A small organisation should be very clear about red lines.

For example:

• do not paste sensitive personal data into public AI tools
• do not upload confidential contracts or board papers without permission
• do not use AI-generated facts without checking them
• do not let AI make decisions about people
• do not present AI-written material as lived experience
• do not use AI outputs in final submissions without human review

This is the part that protects trust.

Keep human review visible

An AI policy should say who is responsible for checking output.

That means:

• checking facts
• checking tone
• checking funder requirements
• checking data and evidence
• checking whether the copy still sounds like the organisation
• checking whether any claim is unfair, exaggerated, or unsupported

AI can draft. People are still accountable.

Make it usable

The best policy is one people can actually follow.

You might create:

• a one-page safe-use checklist
• a traffic light system for use cases
• example prompts
• a review checklist
• a short note for trustees
• a process for approving new tools

This is where an AI Policy and Guidance Starter can help. The aim is to create something clear enough for staff, freelancers, and trustees to use in real work.

Want help writing your AI policy?

ArtsFunded Training supports arts, culture, and charity teams with practical AI workshops, workflow audits, and policy starters.

If your organisation is already using AI informally, a simple policy is one of the best ways to make that use safer and more consistent.