Template Updated October 2026

Sample AI policy for your employee handbook

Employees are already using ChatGPT and other AI tools at work, with or without a policy. A short, clear policy tells them which tools are approved, what information must never go into them, and who checks the output. Copy the template below into your handbook and replace the parts in [brackets].

Start Free Trial All guides
Free to copyPlain-language template with 13 sections
Covers the real risksConfidential data, wrong answers, decisions about people
Fill in the bracketsApproved tools, contacts and review dates are yours to set
Not legal adviceHave it reviewed against the laws where you operate
Template

AI acceptable use policy template

Copy it, then replace everything in [brackets].

AI acceptable use policy 
[COMPANY NAME] ACCEPTABLE USE OF AI TOOLS POLICY
Version [1.0] - Effective [DATE] - Owner: [HR / IT / LEGAL CONTACT]

1. PURPOSE
This policy explains how employees may use artificial intelligence (AI) tools at work, so that we get the benefits of AI while protecting our customers, our colleagues and our company information.

2. SCOPE
This policy applies to all employees, contractors and temporary staff who use AI tools for [COMPANY NAME] work, on company or personal devices. "AI tools" means any software that generates text, images, code, audio or analysis, including chat assistants (for example ChatGPT, Gemini, Microsoft Copilot), AI features inside other software, and AI bots in Microsoft Teams or Google Chat.

3. APPROVED TOOLS
3.1 You may use the AI tools listed at [LINK TO APPROVED TOOLS LIST] for work.
3.2 Approved tools have been reviewed by [IT / SECURITY] for data protection, data retention and security.
3.3 To request a new tool, contact [IT CONTACT]. Do not use an unapproved tool for work tasks while the request is pending.
3.4 Use your company account for approved tools. Do not use personal accounts for company work.

4. INFORMATION YOU MUST NOT ENTER
Unless a tool is approved for that type of data, do not enter:
- personal data about customers, applicants or colleagues (names with contact details, ID numbers, health information, pay, performance records);
- confidential business information (financials, contracts, pricing, plans, source code, passwords, API keys);
- information that a customer or partner gave us under a confidentiality agreement.

If you are not sure whether information is confidential, treat it as confidential and ask [CONTACT].

5. CHECK THE OUTPUT
5.1 AI tools can be wrong, out of date or biased, and can state wrong answers with confidence. You are responsible for any work you produce with AI.
5.2 Check facts, figures, quotes, citations and code before you use or share them.
5.3 Do not rely on AI output alone for legal, medical, financial or safety decisions.

6. DECISIONS ABOUT PEOPLE
AI tools must not be the only basis for decisions about hiring, promotion, pay, discipline or dismissal. A person must review any AI-assisted input to these decisions, and [HR] must approve any AI tool used in recruiting or performance management before it is used.

7. TRANSPARENCY
7.1 Tell the recipient when content you send to customers or the public was mostly generated by AI, where [COMPANY NAME] guidelines or law require it.
7.2 Do not use AI to impersonate a real person or create misleading content.

8. INTELLECTUAL PROPERTY
8.1 Do not enter third-party content you do not have the right to use.
8.2 Check AI output for copied material before publishing it.
8.3 Work you create with AI tools for [COMPANY NAME] belongs to [COMPANY NAME], as set out in your employment contract.

9. PROHIBITED USES
Do not use AI tools to:
- break the law or any company policy;
- harass, discriminate against or monitor colleagues;
- get around security controls;
- create content that is offensive or harmful.

10. SECURITY AND INCIDENTS
10.1 Protect your AI tool accounts with multi-factor authentication where available.
10.2 If you entered information you should not have, or you notice an AI tool behaving unexpectedly, report it to [IT / SECURITY CONTACT] immediately. Early reporting is expected and will not by itself lead to disciplinary action.

11. TRAINING
Employees who use AI tools for work must complete [NAME OF TRAINING] within [NUMBER] days of starting to use them.

12. BREACHES
Breaking this policy may lead to disciplinary action in line with the [DISCIPLINARY POLICY].

13. QUESTIONS AND REVIEW
Questions about this policy go to [CONTACT]. [OWNER] reviews this policy at least every [6 / 12] months, because AI tools and laws change quickly.
Not legal advice

This template is a starting point, not legal advice. Laws differ by country and change often. In the EU, for example, the AI Act already requires employers that use AI systems to make sure staff have enough AI literacy. Have HR and legal review the final version.

How To Use It

How to adapt the policy for your company

  1. 1

    Make the approved tools list first

    The policy points to a list instead of naming tools, so you can update the list without re-approving the policy. Start with the tools people already use and decide which are allowed for which data.

  2. 2

    Decide who owns it

    Name one owner, usually HR with IT. Questions go to them, and they review the policy every 6 to 12 months.

  3. 3

    Match it to your data rules

    Section 4 should use the same terms as your data protection and confidentiality policies, so employees do not get two definitions of "confidential".

  4. 4

    Tell people, then make it easy to ask

    A policy nobody reads changes nothing. Announce it, link it from the handbook, and make the answers easy to find where people already work.

What To Include

What every AI policy for employees should cover

  • Which AI tools are approved, and how to request a new one.
  • What information must never be entered into an unapproved tool.
  • That employees are responsible for checking AI output before using it.
  • That AI is never the only basis for decisions about hiring, pay or discipline.
  • When to tell customers or the public that content was made with AI.
  • Who to tell if confidential data was entered by mistake, without fear of blame.
  • Training, consequences for breaches, and a review date.
Answering Policy Questions

Let employees ask about the policy in chat

After a new policy goes out, HR gets the same few questions: "Can I use ChatGPT to draft this email?", "Is our customer list confidential?", "Which tools are approved?". Add the policy to an HR bot in Microsoft Teams or Google Chat and employees get an answer that quotes the policy section, at any hour.

ChatGridAI does this from your own documents, on your own OpenAI account, and it is itself a tool you can put on the approved list: documents and questions are processed in your company's OpenAI account, not shared with other customers.

FAQ

Common questions

If employees use AI tools at work, yes. Without one, people guess what is allowed, and confidential information ends up in personal AI accounts. A short policy with an approved tools list and clear data rules prevents most problems.
Approved tools and how to request new ones, information that must not be entered, responsibility for checking output, rules for decisions about people, transparency, intellectual property, prohibited uses, incident reporting, training and a review date.
That is your company's decision. Many companies allow approved AI tools with company accounts and ban entering personal or confidential data into tools that are not approved for it. The template above follows that approach.
No. It is a plain-language starting point. Laws differ by country and change often, so have HR and legal counsel review the final version before you publish it.
At least every 6 to 12 months. AI tools, vendor terms and laws are changing quickly, and the approved tools list may need updates more often than that.

Answer policy questions in chat

Put your handbook and AI policy into an HR bot for Teams and Google Chat. Answers cite the section they came from.

Start Free Trial Read the setup guide

$5/seat/month - 14-day free trial with 10 seats - no credit card required