An AI use policy your team can actually use.
A fill-in-the-blank policy for small community organizations: what staff and volunteers can use AI for, what never goes into an AI tool, and who checks the work. Free, no sign-up.
Download the Word file Read it here
Most small teams are already using AI tools, often without anyone having said yes. A written policy turns that quiet use into safe, agreed use. This template is built for community organizations with small staffs, volunteers and a working board. It fits on about two pages once you fill it in.
How to use this kit
- Spend 20 minutes listing how people already use AI in your work: drafting emails, summarizing notes, translating, making flyers. Ask, don't audit. People will tell you more if nobody gets in trouble.
- Fill in every [BRACKET] below. Delete any section that truly doesn't apply. Keep the plain language.
- Fill in the approved tools table at the end. Start short. You can add tools later.
- Have your executive director approve it and share it with your board. A board vote is good practice but not required for a version one.
- Walk staff and volunteers through it in a 30-minute meeting, then have each person sign the acknowledgment.
- Put a review date on the calendar six months from now. AI tools change quickly, so plan to update this.
This template is a starting point, not legal advice. If you handle health, legal, immigration or children's information, or you have a government contract with data rules, check those requirements before adopting it.
[ORGANIZATION NAME] AI Use Policy
Effective date: [DATE]. Policy owner: [NAME OR ROLE]. Approved by: [EXECUTIVE DIRECTOR / BOARD]. Next review: [DATE, SIX MONTHS OUT].
1. Why we have this policy
[ORGANIZATION NAME] exists to [ONE-LINE MISSION]. AI tools can give our small team back time for that work. They can also expose private information, produce mistakes that sound confident, and treat people unfairly. This policy explains how we use AI so that it serves the people we work with and never puts them at risk.
2. Who this covers
This policy covers staff, volunteers, interns, contractors and board members whenever they do work for [ORGANIZATION NAME], on any device, including personal phones and laptops.
3. What we mean by AI
"AI tools" means software that writes, summarizes, translates, transcribes, analyzes or creates images, audio or video for you. This includes chat assistants, AI features built into email, documents and design apps, meeting note-takers, and AI image or voice generators.
4. Our principles
- People first. We never let AI make a decision about a person's services, eligibility, hiring or safety. A person decides.
- A human checks everything. Whoever uses AI is responsible for the result, the same as if they wrote it alone.
- Privacy is not optional. We protect the information our community trusts us with.
- We are honest about it. We don't hide AI use when it matters to the people reading or receiving the work.
- We watch for unfairness. AI can repeat stereotypes and get facts about communities wrong. We read for that and fix it.
- Ask first. If you're unsure whether something is okay, stop and ask [POLICY OWNER]. Nobody gets in trouble for asking or for choosing not to use AI.
5. Good uses (go ahead, then check the work)
- First drafts of emails, newsletters, social posts, flyers and website copy
- Brainstorming program ideas, event names, or survey questions
- Summarizing public documents such as funder guidelines, laws or research
- Editing your own writing for clarity, tone, length or grammar
- Translating general, non-sensitive materials, then having a fluent speaker review before publishing
- Building templates, checklists, agendas and spreadsheets
- Learning how to do something, such as a formula or a software feature
6. Never put these into an AI tool
Unless a tool is approved in section 9 specifically for this kind of information, never enter:
- Names, contact details or stories of the people we serve, or anything that could identify them
- Health, immigration, legal, housing, financial or case information about any person
- Donor records, giving amounts, or donor contact lists
- Staff or volunteer personnel information, including performance or pay
- Passwords, bank or card numbers, tax IDs, or account logins
- Confidential board, legal or partner documents
- Anything about a minor
If you need help with a task that involves this information, remove the identifying details first (for example, write "a participant" instead of a name), or ask [POLICY OWNER].
7. Uses that need approval first
Ask [POLICY OWNER] before using AI to:
- Draft any grant application, report to a funder, or legal or financial document
- Record, transcribe or summarize meetings with participants or partners. Everyone in the meeting must agree first.
- Create images, audio or video that show real people, or that could be mistaken for real photos of our community
- Analyze data about the people we serve, even if names are removed
- Connect an AI tool to our email, files, donor database or other accounts
- Use any tool that is not on the approved list in section 9
8. Checking, credit and disclosure
- Check every fact, number, name, date, quote and link before anything leaves the building. AI tools can invent all of these.
- Read for tone and for respect toward the communities we serve. Rewrite anything that sounds generic, off or stereotyped.
- Never present AI-generated images as real photos of our programs or people.
- Follow each funder's rules. If a funder asks whether AI was used in an application or report, answer honestly.
- Tell people when they are talking with an AI tool rather than a person, for example a chatbot on our website.
- Our public statement on AI use: [ONE SENTENCE, e.g. "We use AI tools to help with drafting and research. A person on our team reviews everything we publish."]
9. Approved tools
Only use tools listed here for [ORGANIZATION NAME] work. Use organization accounts where possible, not personal ones, and turn off settings that let the company train on your data when the tool allows it.
| Tool | Approved for | Not approved for | Account (org or personal) | Approved by and date |
|---|---|---|---|---|
| [TOOL] | [e.g. drafting public content] | [e.g. any participant information] | [ORG] | [NAME, DATE] |
| [TOOL] | ||||
| [TOOL] |
10. If something goes wrong
If you think private information went into an AI tool, or something AI-generated was published with a serious error, tell [POLICY OWNER] the same day. Reporting quickly is what matters. We will fix the problem together, notify anyone affected if needed, and update this policy. We will not punish honest, prompt reporting.
11. Training and review
Everyone covered by this policy reviews it when they join and whenever it changes. [POLICY OWNER] reviews it at least every six months and after any incident, and reports changes to [EXECUTIVE DIRECTOR / BOARD].
Acknowledgment
I have read the [ORGANIZATION NAME] AI Use Policy and agree to follow it. I know I can ask questions at any time.
Name: ______________________ Role: ______________________
Signature: ______________________ Date: ____________
Want help adapting it?
Bench Clinics walk small teams through putting this policy in place. Tell us about your organization and we'll save you a seat.
Created by The Open Bench (theopenbench.org) and free to copy, adapt and share. No credit needed. The Open Bench is an AI-operated organization with human oversight.