What is an AI acceptable use policy?

An internal document that says which AI tools people may use, for what, with which information, and who checks the results. It turns informal habits into rules everyone can follow and an auditor can read.

Who is the AI Policy Builder for?

Companies that already use AI tools at work and have no AI governance specialist: small and mid-sized businesses, agencies and consultancies, and AI startups. One set of questions covers all of them; your answers change what the policy contains.

Does the policy follow the EU AI Act?

The clause library uses the EU as its baseline. Locked sections draw on the EU AI Act (prohibited practices, transparency, AI literacy), the GDPR and the ISO/IEC 42001 policy structure. If you operate outside the EU, a flag tells you which rules the baseline does not cover.

How is it different from a template or a chatbot?

A template leaves you to decide what applies. A chatbot writes new text every time. The builder selects sections by rules, keeps legal wording fixed, shows why each section is there and flags what needs a decision. AI only adapts the wording of a few sections, and only if you ask it to.

Is the generated policy legal advice?

No. It is a starting point for your own policy. Clause library v0.1 is a working draft that the AIEI Legal Committee has not yet reviewed, and the Legal flags show where to take advice.

The AI Policy Builder is included in AIEI membership for individuals and organisations. Become a member, then log in and open this page.

AI use policy generator for companies using AI

Answer 19 questions about how your team uses AI. Get a policy built from AIEI’s clause library, with every gap flagged.

Become a Member Log in

AIEI-POL-AUP Example company

AI Acceptable Use Policy

Northwind Studio Ltd, version 1.0

7Information and confidentialityLocked wording

Never enter passwords, access keys or other credentials, and do not paste in more information than the task needs.

9Client work and IPAdded: you work for clients

Before using AI on client work, check the client agreement. Where a client has restricted AI use, the client’s terms take precedence.

Adapted with AI For a design studio, this includes moodboards and drafts shared for approval.

4Approved tools and accounts

Tools are used through the business accounts Northwind Studio provides. Personal accounts are not used for work.

ConflictWork data and personal accountsYour team mostly uses personal accounts. Resolve before you publish.

Example company. Your policy is built from your own answers.

Six steps, about 15 minutes

Every question is closed-choice and describes how you work today. Answers stay saved in your browser, so you can stop and come back.

1
Organisation

Size, sector, regions, client work and the tone you want.
6 questions
2
AI use

Tools, accounts, what AI is used for and what reaches the public.
4 questions
3
Data

What goes into AI tools and what your vendors may do with it.
3 questions
4
Oversight

Owner, approver, incident contact, checks, training, review cycle.
6 questions
5
Review

Read the draft, see why each section is there, resolve the flags.
Flags listed
6
Export

Finalise the wording and download the policy.
Word or PDF

Needs attention 3 items

Conflict Work data and personal accounts
You said work information may reach AI tools and that some people use personal or free accounts. Make sure company accounts exist before you publish.
Change answer
Legal Possible high-risk use
AI in recruitment or performance decisions may be high-risk under the EU AI Act (Annex III). Have that section checked by counsel.
Change answer
Check Vendor training on your inputs
Check each tool’s settings. Inputs used to train a vendor’s models leave your control; business plans usually let you switch this off.
Change answer

Your policy: 17 sections + annex 3

Flags before you publish

Conflicts, legal review points and missing facts are listed in one place, each linked to the answer behind it. Nothing is hidden in the text.

Your answers stay in your browser
Company name, owners and answers are never stored on our server.
AI only when you ask
“Adapt with AI” sends the adaptable sections and your answer labels, nothing else.
Locked sections never leave
Wording with legal weight is never sent to AI and stays as the library wrote it.

FAQ

What is an AI acceptable use policy?

An internal document that says which AI tools people may use, for what, with which information, and who checks the results. It turns informal habits into rules everyone can follow and an auditor can read.
Who is the AI Policy Builder for?

Companies that already use AI tools at work and have no AI governance specialist: small and mid-sized businesses, agencies and consultancies, and AI startups. One set of questions covers all of them; your answers change what the policy contains.
Does the policy follow the EU AI Act?

The clause library uses the EU as its baseline. Locked sections draw on the EU AI Act (prohibited practices, transparency, AI literacy), the GDPR and the ISO/IEC 42001 policy structure. If you operate outside the EU, a flag tells you which rules the baseline does not cover.
How is it different from a template or a chatbot?

A template leaves you to decide what applies. A chatbot writes new text every time. The builder selects sections by rules, keeps legal wording fixed, shows why each section is there and flags what needs a decision. AI only adapts the wording of a few sections, and only if you ask it to.
Is the generated policy legal advice?

No. It is a starting point for your own policy. Clause library v0.1 is a working draft that the AIEI Legal Committee has not yet reviewed, and the Legal flags show where to take advice.
How do I get access?

The AI Policy Builder is included in AIEI membership for individuals and organisations. Become a member, then log in and open this page.