[ORGANIZATION NAME] AI USE POLICY
========================================

Organization name: [fill in]
Policy version: [fill in]
Date adopted: [fill in]
Adopted by (board vote / leadership): [fill in]
Next review date: [fill in]
Policy owner (name and role): [fill in]

(Note: How to use this template: fill in every box and tick every checkbox that applies. Delete anything that does not fit your organization. When it is finished, have your board or leadership adopt it, then give the whole document to your AI tool as standing instructions (see Part 6). Keep the wording plain. Your AI tool and your newest volunteer should both be able to follow it.)


PART 1. WHO WE ARE
------------------

(Note: The AI reads this first. Use your exact, approved wording. Anything you leave out, the AI may try to fill in by guessing.)

Legal name: [fill in]
Name we use in public: [fill in]
Year founded: [fill in]
City / service area: [fill in]
Website: [fill in]
Type of organization (nonprofit, church, ministry, etc.): [fill in]

Mission statement (exact wording):
[fill in]

Who we serve (describe them the way you would want them to hear it):
[fill in]

Program or ministry | What it does, in one sentence
[fill in] | [fill in]
[fill in] | [fill in]
[fill in] | [fill in]

What makes us different:
[fill in]


Approved facts

(Note: Only facts listed here may be used in anything AI helps write. Every number must have a source. Update this list at least once a year.)

Fact or number | Where it comes from | Last checked
[fill in] | [fill in] | [fill in]
[fill in] | [fill in] | [fill in]
[fill in] | [fill in] | [fill in]


PART 2. OUR BRAND STYLE
-----------------------

(Note: This is what makes the AI sound like you instead of like everyone else.)

Three words that describe how we sound: [fill in]
Three words we never want to sound like: [fill in]
Reading level / tone (e.g. warm, plain, 8th-grade reading level): [fill in]
Preferred length (e.g. social posts under 60 words): [fill in]

We say | We never say
[fill in] | [fill in]
[fill in] | [fill in]
[fill in] | [fill in]

How we refer to the people we serve (and words to avoid):
[fill in]

Spelling of our name, programs and leaders' titles: [fill in]
Capitalization and punctuation rules: [fill in]
Who signs our letters and emails: [fill in]
Brand colors (name and hex code): [fill in]
Fonts: [fill in]
Logo rules (where it goes, what never to do with it): [fill in]

A paragraph of our writing we love (paste a real example so the AI can match it):
[fill in]


PART 3. RULES FOR USING AI
--------------------------


3.1 Approved tools

(Note: List every AI tool people may use for our work. Use organization accounts where possible, so our work and history stay with us when someone leaves.)

Tool | Account owner | Who may use it
[fill in] | [fill in] | [fill in]
[fill in] | [fill in] | [fill in]
[fill in] | [fill in] | [fill in]

[x] Only the tools listed above may be used for our work.
[x] Personal AI accounts may not be used for our confidential work.
[ ] Data-sharing / model-training settings have been turned off where the tool allows it.


3.2 What AI may help with

[x] First drafts of social posts, emails and newsletters
[x] Brainstorming ideas, headlines and titles
[x] Fixing grammar, spelling and clarity
[x] Summarizing public documents or our own non-confidential documents
[ ] Drafting translations (reviewed by a fluent speaker before use)
[ ] Drafting outlines for grant applications and reports
[ ] Other: [fill in]


3.3 What needs extra approval before AI is used

(Note: Write the name or role of the person who must approve each one.)

Grant applications and grant reports: approved by [fill in]
Anything with numbers, financial figures or outcomes: approved by [fill in]
Press releases and statements to media: approved by [fill in]
Crisis, safety or sensitive communications: approved by [fill in]
Anything published under a leader's name: approved by [fill in]
AI-generated images or video: approved by [fill in]


3.4 Human review

A person reads and approves everything AI helps create before it is sent, posted, printed or submitted. The person who publishes it is responsible for it, not the AI tool.

Final reviewer for public content: [fill in]
Final reviewer for donor and funder content: [fill in]


3.5 Checking facts

Every name, number, date, quote, statistic and source in AI-assisted work is checked against our own records or the original source before use. If it cannot be checked, it is removed.


3.6 Being honest about AI

[x] We never present AI-generated images as real photos of our people, events or programs.
[ ] We say when content is substantially AI-generated, where the law, platform or funder requires it.
[ ] Other disclosure rule: [fill in]


3.7 When something goes wrong

(Note: Mistakes will happen. What matters is that people report them quickly and without fear.)

Report AI mistakes or data leaks to: [fill in]
Report within (hours): [fill in]
Who decides on corrections or public response: [fill in]


3.8 Training and sign-off

Everyone who uses AI for our work (staff, volunteers, board, contractors) reads this policy and signs the acknowledgment in Part 7 before using AI for us.


PART 4. GOVERNANCE (FROM OUR BY-LAWS)
-------------------------------------

(Note: Your by-laws already say who decides what. This part makes sure AI follows the same lines of authority. Copy the relevant sections, with article and section numbers.)

This policy is adopted under by-laws Article / Section: [fill in]
Who may speak publicly for the organization: [fill in]
Who may sign contracts and approve spending (including AI subscriptions): [fill in]
Who may approve grant submissions: [fill in]

Confidentiality rules that apply (from by-laws, handbook or other policies):
[fill in]

Conflict-of-interest rules that apply:
[fill in]

Record-keeping rules that apply:
[fill in]

Other by-law sections AI-assisted work must respect (paste them here):
[fill in]


PART 5. WHAT AI MUST NEVER DO
-----------------------------

(Note: These are the non-negotiable lines. They apply to every tool, every person and every project. Keep the ones that apply, add your own, and do not water them down.)

1. Never receive personal information about the people we serve, donors, staff or volunteers: names, addresses, phone numbers, case notes, health, addiction, criminal history, immigration status, children's information or anything that could identify them.
2. Never invent facts, statistics, quotes, testimonials, outcomes or stories.
3. Never create a story about a person we serve that did not happen, or combine several people into one story without clearly saying so.
4. Never create images of real people we serve, or images that could be mistaken for real photos of our work.
5. Never make decisions about people: who receives services, who is hired, screened, disciplined or let go.
6. Never give legal, medical, mental-health, tax or financial advice on our behalf.
7. Never speak directly to the public, the people we serve, or anyone in crisis without a person supervising.
8. Never send, post, print or submit anything without human review.
9. Never produce figures for grant reports, financial statements or government filings without checking them against our records.
10. Never receive passwords, bank or card details, donor giving records or non-public financials.
11. Never write as a real person (our pastor, director, a client, a donor) without that person's approval.
12. Never use language that stereotypes, shames or talks down to the people we serve.
13. Never make promises for us: services, outcomes, deadlines, or whether a gift is tax-deductible.

Our additional NEVER rules:
[fill in]


PART 6. INSTRUCTIONS FOR THE AI
-------------------------------

(Note: Paste this whole document into your AI tool's standing instructions (for example a ChatGPT Project or Custom Instructions, a Claude Project, a Gemini Gem or Microsoft Copilot notebook). The block below tells the AI how to use it.)

You are assisting [Organization name]. Everything in this document is a binding rule for you.
1. Write in the voice described in Part 2. Use only the words and terms we approve.
2. Use only facts from the Approved Facts list in Part 1. If you need a fact that is not there, write [VERIFY] instead of guessing.
3. If a request would break any rule in Part 5, refuse, name the rule, and suggest a safe alternative.
4. If someone gives you personal information covered by Part 5, do not use it. Tell them to remove it and use a placeholder like [NAME].
5. Everything you write is a draft for a person to review. Never say it is final.
6. Follow the approval and governance rules in Parts 3 and 4. If a task needs approval, remind the user who must approve it.
7. If you are unsure whether something is allowed, ask before doing it.


PART 7. ADOPTION AND ACKNOWLEDGMENT
-----------------------------------


Adoption

Adopted by: [fill in]
Date: [fill in]
Recorded in minutes of (meeting and date): [fill in]
Signature: [fill in]


Acknowledgment

I have read this AI Use Policy, I understand it, and I will follow it whenever I use AI for this organization's work.

Name | Role | Signature | Date
[fill in] | [fill in] | [fill in] | [fill in]
[fill in] | [fill in] | [fill in] | [fill in]
[fill in] | [fill in] | [fill in] | [fill in]

(Note: This template is a starting point, not legal advice. Have your board review it, and where your work involves confidential, health, legal or children's information, ask a qualified attorney to review it too.)
