Writing an AI use policy for your membership

Rosa runs a nutrition coaching membership with two part-time coaches and a virtual assistant. One coach pastes client food diaries into a free AI app to spot patterns. The other won't touch AI at all. The assistant drafts every newsletter with it and sends them straight out. Nobody is being careless on purpose. There are simply no shared rules, so everyone has invented their own.

An AI use policy fixes that. It is a short, plain document that says how your business uses AI: which tools, for what, with which data, and who checks the results. It is not a legal contract. It is a working agreement that protects your members, your reputation and the people who help you, and it is worth writing even if the only person who reads it is you.

Know who the policy is for

Your AI use policy covers the people who work on your membership: you, staff, contractors, moderators, guest instructors and anyone with admin access. It is separate from any rules you set for how members themselves use AI in your community or courses; those need their own guidance.

You can also publish a short, member-friendly version. Many owners turn the key points into the public statement described in telling members when you use AI, so what you promise publicly matches what your team actually does.

What your policy should cover

Keep it to one or two pages. Most membership businesses need these sections:

1. Approved tools: which AI tools and which accounts may be used for membership work, and who approves new ones. If you are still choosing, see choosing your first AI assistant.
2. Data rules: what may never go into an AI tool, what may go in only after removing identifying details, and what is fine to use. Protecting member privacy when you use AI tools has a three-group approach you can copy.
3. Good uses: tasks where AI is encouraged, such as first drafts, summaries, formatting and brainstorming.
4. Off-limits uses: things AI must never do on its own, such as deciding refunds, removing members, approving applications, writing testimonials or imitating a real person.
5. Review and sign-off: who checks what before it reaches members, and what they check for.
6. Disclosure: when and how you tell members that AI was involved.
7. Other people's work: no pasting in content you do not have rights to, and care with images that imitate a named artist or brand.
8. When something goes wrong: who to tell and how quickly.
9. Ownership and review date: who maintains the policy and when it will next be reviewed.

Sample policy wording

Here is wording you can adapt. Replace the bracketed parts and delete anything that does not fit how you work:

AI use at [membership name]

We use AI tools to save time on drafting, summarizing and formatting so we can spend more time with members. We use only these tools, through business accounts managed by [owner name]: [approved tools]. Personal AI accounts are not used for membership work.

We never enter passwords, payment details, health information, information about children or anything a member shared in confidence. Before using any member message or file, we remove names, contact details and anything else that could identify someone.

AI may help draft content, support replies and summaries. AI does not make decisions about refunds, cancellations, account removals, applications or disputes; a person makes those decisions. Every AI-assisted item is reviewed by [role] before it reaches a member, checking facts, tone, promises and privacy.

If you think something has gone wrong, such as member data entered into the wrong tool or an incorrect answer sent to a member, tell [owner name] the same day. This policy is reviewed every [period] by [owner name].

Let AI help you draft it, then make it yours

An assistant can turn your rough notes into a tidy first draft. Give it your real practices rather than asking it to invent a policy, because an invented one will describe rules you do not follow:

I run [type of membership] with [who helps you]. Here are my notes on how we use AI today and the rules I want: [your notes]. Turn these into a plain-language AI use policy of no more than [length], with short sections for approved tools, data rules, good uses, off-limits uses, review, disclosure and incidents. Use only the rules in my notes. Where my notes are unclear or missing something important, list questions for me at the end instead of making up an answer.

Then read every line as if you were a new contractor. Is each rule clear enough to follow without asking you? Does anything promise more than you can deliver? If any section touches on legal obligations, such as data protection or employment terms, have someone qualified review it.

Roll it out so people follow it

A policy that sits in a shared folder changes nothing. Walk each person through it, using real examples from your membership: here is a support email, here is how to strip out the details, here is who reviews the draft. Ask what feels unworkable and adjust. Rules people helped shape are rules they keep.

Build the policy into your routines. Add it to contractor onboarding, and when someone leaves, remove their access to shared AI workspaces along with everything else; the guide to offboarding staff and contractors securely covers that wider process. Revisit the policy whenever you add a new tool or a new kind of AI task.

First steps

• List the AI tools and uses already happening across your membership.
• Draft your policy from the sections above, or from the prompt and your notes.
• Review the draft line by line and remove anything you would not actually enforce.
• Walk everyone through it and collect their questions.
• Publish a short member-facing summary that matches the policy.
• Put a review date in your calendar.