Launch offer20% off every Pro version until 30 November, applied at checkoutSee the pricesHoliday offer15% off every Pro version until 31 December, applied at checkoutSee the prices
MutewareAdd to Chrome

AI policy template for your company

A one page policy you can adopt as it is: what colleagues may use AI tools for, what never gets pasted into them and who decides. Copy it, replace the words in square brackets and share it with your team.

[Company name]: policy for using AI tools

Version of [date]. Owner: [name or role].

1. Why we have this policy

AI tools such as ChatGPT, Claude, Gemini and Copilot help us write, summarise, translate and program faster, and we want everyone to use them well. This policy says how to do that without putting customers, colleagues or the company at risk.

2. Tools we use

  • Approved for work: [the tools and plans your company has chosen].
  • For work you use the company account, not a private account.
  • A new AI tool is approved by [role] before anyone uses it with company information.

3. What never goes into an AI tool

  • Passwords, API keys, access tokens and private keys.
  • Sensitive personal data: health, religion, ethnic origin, political views, trade union membership, sexual orientation, biometric and genetic data, and criminal records.
  • National identification numbers, copies of identity documents, and bank account and card numbers.
  • Information that we promised a customer or partner to keep confidential, unless the contract allows this use.
  • [Your own additions, for example unpublished financial figures or the source code of a certain product.]

4. Personal data

  • Use as little personal data as the task needs. Most tasks work just as well when names and numbers are replaced by placeholders.
  • Replace names, email addresses, phone numbers and customer numbers before you paste, and put them back in the answer. We use [tool] for this.
  • AI tools do not decide about people. For hiring, reviews, credit and similar decisions a person looks at the facts and decides.

5. Check the result

  • AI tools make mistakes and state them with confidence. You remain responsible for what you send, publish or build.
  • Check facts, figures, quotes and sources. Test code before it is used.
  • Do not copy long passages from an answer into our own publications without checking where they come from.

6. Be open about it

  • People who talk to a chatbot of ours are told that it is an AI system and not a person.
  • Images, audio and video made with AI that could be mistaken for real are labelled as such.
  • [Your rule for mentioning the use of AI in work for customers.]

7. Settings and connections

  • Where a tool lets you choose, switch off the use of our content for training its models. [How to do this in the tools we use.]
  • Do not connect an AI tool to mailboxes, shared drives or other company systems without approval from [role].
  • Browser extensions and plug-ins for AI tools are approved by [role] before you install them.

8. Knowledge

Everyone who uses AI tools for work follows [the introduction or training], so that they know what these tools can and cannot do. New colleagues do this in their first month.

9. When something goes wrong

  • Pasted something that should not have gone in? Report it right away to [contact]. Reporting fast is always the right thing to do, and nobody is blamed for reporting.
  • If it was a password or a key, change it immediately.
  • [Contact] decides whether it is a data breach that has to be reported.

10. Review

[Role] reviews this policy every [six] months, and whenever we start using a new AI tool.

This template is general information and not legal advice. You are free to use and change it.

How to introduce it

  1. Fill in the square brackets and cut what does not fit your company. Keep it short: a policy nobody reads protects nobody.
  2. Walk through it in a team meeting and show what it looks like in practice, for example how to scrub a text before you paste it.
  3. Put it where everyone can find it, and agree on who answers questions.

Questions

Can I use this policy as it is?

Yes. It is free to copy, change and share inside your organisation, and you do not have to mention us. Replace the words in square brackets, remove what does not fit and add your own rules.

Is this legal advice?

No. It is a practical starting point written for small and medium sized companies. It does not replace advice about your own situation. If you process a lot of personal data or work in a regulated sector, have your data protection officer or a lawyer read it.

What does the law ask?

In the EU the GDPR applies as soon as personal data goes into an AI tool, which usually means you need a data processing agreement with the provider and a good reason to share the data at all. The EU AI Act has also asked organisations that use AI systems, since 2 February 2025, to make sure their staff know enough about AI to use it responsibly. Section 8 of the template covers that. More background is in the GDPR checklist for AI chatbots at work.

How do we make rule 4 easy to follow?

By making the safe way the fast way. Scrub for AI replaces personal data and secrets with placeholders in the browser and puts them back in the answer, so nobody has to do it by hand.

Make the safe way the fast way

Scrub for AI replaces personal data and secrets with placeholders before they reach an AI chat, and puts them back in the answer. Free for everyone, with yearly plans for teams.