Church AI Policy Template

Who this is for: the person putting together an AI policy for church council or leadership team approval, after being asked to look into AI.

How to Use This Template

This is a ready-to-adopt policy document. You can take it to your next church council, deacons’ or trustees’ meeting as it stands, or adapt it to suit your context.

To adopt it:

1. Read through the whole document with your leadership team
2. Amend anything that doesn’t fit your church’s situation (see the notes in square brackets)
3. Add your church name and the date
4. Put it to your governing body for formal approval
5. Share it with staff and volunteers
6. Diarise an annual review

The language is deliberately denomination-neutral. Where governance structures differ, we’ve included options so you can choose the term that fits your tradition (e.g. church council, PCC, deacons, elders, trustees).

---

[Church Name] - Artificial Intelligence (AI) Policy

Adopted by: [PCC / Church Council / Deacons / Elders / Trustees - delete as appropriate] Date adopted: [Date] Review date: [Date - we recommend 12 months from adoption]

---

1. Purpose

This policy sets out how [Church Name] will use artificial intelligence tools responsibly, safely and in line with our Christian values. It applies to all AI tools, including but not limited to ChatGPT, Claude, Microsoft Copilot, Google Gemini, Midjourney and any AI features built into software we already use (such as Canva, Mailchimp or Google Workspace).

We believe AI can be a helpful tool for church administration and communication. We also recognise that it requires careful, thoughtful use - particularly where personal data, pastoral sensitivity or theological integrity are involved.

---

2. Scope

This policy applies to:

• All paid staff
• All volunteers acting on behalf of the church
• Anyone creating content, communications or documents for church use
• Anyone using church accounts or devices

It covers the use of AI tools for any church-related purpose, whether on church devices or personal devices.

---

3. Approved Uses

AI tools may be used for the following purposes, subject to the safeguards in this policy:

Administration

• Drafting agendas, minutes and reports
• Summarising long documents
• Drafting emails and letters (non-pastoral)
• Creating to-do lists and action plans
• Formatting and proofreading documents

Communications

• Drafting newsletter articles
• Writing social media posts
• Creating promotional text for events
• Suggesting subject lines for emails
• Generating image ideas or descriptions

Sermon and Teaching Research

• Exploring Bible passages and background context
• Researching historical or theological topics
• Generating discussion questions for small groups
• Brainstorming sermon illustrations
• Summarising commentaries or articles

Training and Development

• Creating training materials for volunteers
• Drafting role descriptions
• Producing induction documents
• Generating quiz or icebreaker questions

---

4. Prohibited Uses

AI tools must not be used for:

Safeguarding

• Making safeguarding decisions or risk assessments
• Drafting safeguarding reports or referrals
• Analysing or processing any safeguarding information
• Safeguarding matters must always be handled by trained personnel following your denomination’s safeguarding procedures.

Where to go for safeguarding advice. Nothing on this site is safeguarding guidance and we do not write any. Your denomination’s safeguarding policy and its safeguarding team are the authority, and they take precedence over anything here. For independent advice, thirtyone:eight runs a helpline for churches on 0303 003 1111; the NSPCC Helpline is 0808 800 5000. Allegations against an adult working with children go to your local authority LADO. If a child or adult is at immediate risk, call 999.

Pastoral Care

• Replacing human pastoral conversations
• Generating pastoral advice to send to individuals
• Diagnosing or advising on mental health, medical or legal matters
• Drafting sensitive pastoral letters without significant human rewriting and review

Theology and Doctrine

• Generating sermons, prayers or theological statements for use without thorough human review and editing
• Presenting AI-generated theological content as the minister’s or preacher’s own work without disclosure to the leadership team
• Using AI to settle doctrinal disputes

Personal and Sensitive Data

• Entering any personal data into free-tier AI tools (see Section 5)
• Entering names, addresses, phone numbers or email addresses
• Entering financial information about individuals (e.g. giving records)
• Entering any information about children or vulnerable adults
• Entering any health information, prayer requests with personal details or pastoral notes

---

5. Data Handling

What you must never enter into any AI tool

• Names or identifying details of church members
• Safeguarding information of any kind
• Financial information about individuals
• Health information or pastoral notes
• Passwords, login details or account numbers
• Any information about children or vulnerable adults

Free vs paid AI tools

Free-tier tools (e.g. free ChatGPT, free Gemini):

• Typically use your inputs to train their models on consumer tiers - including the paid ones. Policies change, and they have changed recently: under terms published on 28 August 2025, with a decision deadline of 8 October 2025, Anthropic trains on consumer Claude conversations (Free, Pro and Max) where the setting is on, with retention of up to five years for chats not opted out. Assume every consumer chatbot trains on what you type unless you have checked its current policy yourself, today.
• The exclusions are generally the business, education, government and API tiers, not the free or personal-paid ones. If you need a genuine no-training guarantee, that is the tier to buy - and it is a different product, not a setting
• Offer weaker data protection
• Should only be used for fully anonymised, non-sensitive tasks (e.g. “Write a social media post about a Harvest Festival” - no personal details)

Business, education and API tiers (e.g. ChatGPT Team/Enterprise, Claude Team, Microsoft 365 Copilot):

• Generally do not use your inputs for training. ⚠️ Personal paid subscriptions such as ChatGPT Plus and Claude Pro are consumer tiers and are not covered by this - see the paragraph above. A genuine no-training guarantee is a different product, not a setting
• Offer stronger data handling commitments
• Are preferred for any church work, though the rules above about personal data still apply

Note: Even paid tools are not a safe place for personal data. The safest approach is to never enter identifiable personal information into any AI tool.

UK GDPR Compliance

Under UK GDPR, your church is a data controller. Using AI tools to process personal data may constitute a data processing activity that needs to be recorded in your Records of Processing Activities (ROPA) and, potentially, assessed via a Data Protection Impact Assessment (DPIA).

If in doubt, consult your church’s data protection lead or contact your denomination’s data protection adviser.

Church of England: Ask your diocesan office who handles data protection for parishes. Methodist: Refer to the Methodist Church’s GDPR guidance at methodist.org.uk. Baptist: Contact your Baptist Union (Great Britain, Scotland or Wales) or your regional association. Independent: Consider consulting the Information Commissioner’s Office (ICO) guidance for small organisations. Any denomination: The ICO runs a helpline for small organisations; the number is on ico.org.uk.

---

6. Quality Assurance

All AI-generated content must be reviewed by a named person before it is used, published or distributed.

Content Type	Reviewer
Social media posts	[Name / Role]
Newsletter articles	[Name / Role]
Sermon material	[Minister / Preacher]
Official letters and reports	[Name / Role]
Training materials	[Name / Role]
Website content	[Name / Role]

Review checklist

Before publishing or using AI-generated content, the reviewer should check:

• Is the content factually accurate?
• Is the theology sound and consistent with our church’s beliefs?
• Does the tone feel authentic and appropriate?
• Has all personal or identifiable information been removed?
• Are any Bible references correct? (AI tools frequently get these wrong)
• Does the content reflect our church’s voice, not a generic AI voice?
• Would you be comfortable if a church member knew this was AI-assisted?

---

7. Transparency

We believe in honest, transparent communication. The following guidelines apply:

When to disclose AI use

• Always disclose to the leadership team if AI has been used to draft sermons, prayers or theological content
• Good practice to disclose in newsletters or on social media if AI has assisted with content creation (e.g. “Written with AI assistance” or similar)
• Not necessary to disclose for routine administrative tasks (formatting, proofreading, summarising)

How to disclose

Keep it simple and honest. Examples:

• “This article was drafted with the help of AI and reviewed by [Name].”
• “AI tools were used to research background for this sermon.”
• A general note in your newsletter: “We sometimes use AI tools to help draft communications. All content is reviewed and approved by [Name/Role].“

---

8. Children and Young People

Additional restrictions apply when AI is used in connection with work involving children and young people:

• No personal data about children or young people may be entered into any AI tool, under any circumstances
• AI-generated content for children’s or youth work must be reviewed by the designated safeguarding lead as well as the content reviewer
• AI tools must not be used to communicate directly with children or young people on behalf of the church
• AI-generated images must not depict identifiable children (note: AI image generators can sometimes produce realistic images of fictional children - these should not be used in church materials)
• Any AI tool used in a session with children or young people must be supervised at all times by an adult who is cleared to work with children under your denomination’s safeguarding policy, which sets the checks required and takes precedence over this template

Note: These restrictions apply regardless of the age of the person using the AI tool. A youth leader aged 20 is subject to the same restrictions as a volunteer aged 70.

---

9. Review Schedule

This policy should be reviewed:

• Annually - at minimum, once every 12 months
• When new AI tools are adopted by the church
• When significant changes occur in AI technology, data protection law or denominational guidance
• After any incident involving AI misuse or data breach

The review should be conducted by [Name / Role] and any changes brought to [PCC / Church Council / Deacons / Elders / Trustees] for approval.

Denominational guidance: Check your denomination’s website annually for updated guidance on AI. This is a fast-moving area and new guidance is being issued regularly.

• Church of England: churchofengland.org - search “artificial intelligence”
• Methodist Church: methodist.org.uk
• Baptist Union: baptist.org.uk (Great Britain), scottishbaptist.com (Scotland) or buw.wales (Wales)
• United Reformed Church: urc.org.uk
• Any denomination: the ICO’s guidance on AI and data protection at ico.org.uk

---

10. Approval and Sign-off

This policy was reviewed and approved by [PCC / Church Council / Deacons / Elders / Trustees] on [Date].

Chair / Moderator / Minister:	_________________________
Print name:	_________________________
Date:	_________________________
Secretary / Clerk:	_________________________
Print name:	_________________________
Date:	_________________________

Next review due: [Date - 12 months from adoption]

---

Appendix: Quick Reference Card

Consider printing this on a single sheet and displaying it near church computers or in the office.

AI in Our Church - Quick Rules

DO:

• Use AI to help draft admin documents, social media posts and newsletters
• Always review AI output before using it
• Keep personal data out of AI tools
• Tell the leadership team if you use AI for sermon preparation
• Check Bible references - AI gets them wrong regularly

DON’T:

• Enter anyone’s name, address or personal details into AI
• Use AI for safeguarding decisions - ever
• Publish AI-generated content without a human review
• Use AI to replace pastoral conversations
• Assume AI output is accurate - always check

NOT SURE? Ask [Name / Role] before proceeding.