Free staff AI use policy template for a small office

A plain-English policy that tells staff which AI tools they can use, what must never go into them, and what to do when something goes wrong. Download it, fill in the brackets and make it yours.

Download the Word file (.docx)

General information only. This is a template, not legal advice. It will not suit every office as it stands. Check it with your own adviser before you use it.

How to use it

1. Replace everything in [square brackets] with your own details, and delete any section that does not apply to your office.
2. Decide which AI tools are approved, and write them in section 3. If you only use a computer in your own office, say so.
3. Check it with your adviser. Privacy law may apply to your business, and your professional body, regulator or insurer may have rules about using AI tools on client information.
4. Give it to your staff, talk it through, and keep a signed copy of the acknowledgement for each person.
5. Review it at least once a year.

The template

Staff AI use policy
[Office name] · Version 1 · Date: [date] · Review date: [date]

1. Why we have this policy

AI tools can save time, but they can get things wrong, and they can expose confidential information if they are used carelessly. This policy explains how we use them at [Office name].

2. Who it applies to

All staff, contractors and anyone else who uses our computers, accounts or documents for work.

3. Approved AI tools

You may only use the AI tools listed here for work: [list the approved tools, for example: our own office AI computer, or the name of any cloud tool you have approved]. Do not use any other AI tool, app or website for work without approval from [name or role].

4. What you may use AI for

Drafting and tidying emails and documents, summarising documents, finding information in our own files, and brainstorming. [Add or remove items to suit your office.]

5. What you must not put into an AI tool

Never enter passwords, login codes or payment card numbers. Do not enter [client names and details, financial information, health information, personal information about staff or other people, anything marked confidential] into any AI tool unless it is on the approved list for that kind of information. If you are unsure, ask [name or role] first.

6. Check everything

AI can be wrong, miss something or make things up, and it can make mistakes with numbers. You are responsible for any work you send out. Check facts, figures, names, dates and sources against the original documents. Where advice or a decision depends on the answer, have a suitably qualified person review it.

7. Be open about it

If you are unsure whether AI should be used for a task, ask. Where a client asks, or where the law or our professional rules require it, we will tell them that AI was used. [Adjust to suit your office.]

8. Records and monitoring

Our approved systems may keep records of the questions asked and the answers given, and [administrators or named roles] may review them for security, quality and to check this policy is followed.

9. Web search and outside services

Some AI tools can search the web. When they do, the words being searched are sent to a search engine. Keep confidential details out of web searches. [Delete this section if you do not use web search.]

10. If something goes wrong

If you put information into the wrong tool, or an AI answer causes a problem, tell [name or role] straight away. We would rather hear about it early, and there is no blame for reporting an honest mistake promptly.

11. Breaches of this policy

Not following this policy may lead to [action, for example: a conversation, retraining, or disciplinary action under our usual process].

12. Review

We will review this policy at least every [12 months], and whenever we change the tools we use.

Acknowledgement

I have read and understood this policy.
Name: ____________________   Signature: ____________________   Date: ____________

Read more: Using AI on your office's own documents and Private AI vs ChatGPT.