AI in the Church AI Church Use: 2026 Report

We asked 208 TGC-aligned church leaders about their attitudes toward and use frequency of 25 different types of AI cases in the local church. We found high consensus among church leaders on what we should and shouldn’t use AI for in our churches. And we found ongoing disagreement primarily on two AI case areas: public/congregation-facing communications and where to draw the line in sermon prep.

Executive Summary

Church leaders have more in common right now on AI than some might think. A consensus has emerged on the clear “yes” and clear “no” for AI usage in a church setting. The “yeses” revolve around internal-facing administrative and operations use cases. The “nos” revolve around pastoral care, counseling, sermon writing/outlining, and chatbots. Opinions vary widely on public/congregation-facing AI-generated content (web copy, social media copy, email copy, graphics) and sermon prep. The coming year or two will likely bring increased usage in the “yes” areas and some of the contested middle-ground use cases.

Read the full report

Enter your name and email and choose the newsletters you’d like to receive. The rest of the report opens right away.

Loading the form…

Five Key Findings

Fig. 1 The Yeses and Nos Are Clear

For each task, the share giving the lowest comfort score (1 of 10) against the share highly comfortable with it, 7 or higher. Seven tasks are refused by a majority.

We found overwhelming consensus to not use AI to write sermons, prayers, or devotionals, clone the pastor, or deploy chatbots.

We found strong consensus that AI is fine to use for organization-facing administration and operational matters like project management, meeting notes and transcripts, assimilation workflows, and volunteer scheduling.

We found strongly diverging opinions surrounding AI assistance in various aspects of sermon preparation and congregation- or public-facing communications.

Fig. 2 The Gray Areas of Comms and Sermon Prep

Each dot is one of 25 tasks, placed by mean comfort (1–10) and how often leaders use AI for it. Color marks the area of ministry. n=208.

  1. The vertical part of this graph indicates increasing comfortability with an AI use case, and the horizontal part of this graph indicates the frequency of usage. The five main areas of this graph are denoted by shading in each area with the respective label.

  2. Embraced

    Each item here is both used fairly regularly with a high degree of comfortability.

  3. Permitted, Not Yet Practiced

    Each item here has a high degree of comfortability but is not yet being widely utilized.

  4. Used Despite Unease

    Sermon research is used frequently despite varying degrees of unease surrounding its usage. Note that website chatbot and pastor-clone bot appear in this quadrant, but these fit more appropriately in the next quadrant of “Holding Back.” They only map here because bots are always on in terms of frequency of usage.

  5. Gray Areas

    We found strong opinions on the areas of congregation- or public-facing communications and where to draw the line of AI assistance in sermon preparation. Some feel great freedom in these two areas, and others’ consciences are constrained. Going forward these will be the two biggest areas of debate.

  6. Holding Back

    Sermon research is used frequently despite a varying degree of unease surrounding its usage. Note that website chatbot and pastor-clone bot appear in this quadrant, but these fit more appropriately in the next quadrant of “Holding Back.” They only map here because bots are always on in terms of frequency of usage.

Fig. 3 Explore What Factors Matter – Part 1

Choose a dimension, then a group, to see its dots move against the full sample. You can look at one group, all groups, or A/B compare any two groups at once. You can look at all categories or any single category of the 25 AI church use cases. Any group with fewer than 8 respondents is hidden.

The map redrawn for one group at a time against the full sample. Groups under 8 respondents are hidden. Differences are tendencies, not single-task certainties.

Fig. 4 Explore What Factors Matter – Part 2

Mean comfortability is scored 1 to 10, and the darker green is more comfortable. Read across a row to see how one task travels between different groups.

Mean comfort (1–10) for every task and every group with at least 8 respondents. Darker is more comfortable. The bottom rows stay pale across every column.

Fig. 5 Permission Exceeds Practice (For Now)

We found several areas where church leaders feel freedom but don’t yet use the emerging technologies. It seems likely that in future surveys these gaps will close a bit as practice begins to match comfortability.

Four Personas

Fig. 6 The Four Personas

One dot is one respondent (208 in all). Bars show each group’s average comfort, 1 to 10, by area of ministry. Groups come from k-means clustering; the boundaries are approximate.

  • Back-Office Pragmatists

    They are comfortable having AI run data, dashboards, systems, operations, and strategic planning. They are uncomfortable using AI in the pulpit or for pastoral care.

  • Study Partners

    They use AI for sermon research, illustrations, and editing as well as administrative uses. They are uncomfortable using AI for pastoral care or to actually outline or write the sermon.

  • Full-Stack Adopters

    They have both the highest comfort and also the most frequent usage of all the personas, with the highest usage and comfortability on all five dimensions. The main place where they draw the line is on AI-written sermons, prayers, and counseling.

  • Red-Line Holders

    They are cautious and skeptical of AI in almost every usage case except for back-office administrative uses, and even then, they are much more reserved than the other three personas.

Take the Quiz

Take this two-minute, ten-question quiz to find out which persona you most closely match.

Church Leaders’ Hopes and Fears

Fig. 7 Church Leaders’ Hopes

174 respondents named at least one hope in their free response. The clear biggest hope is time back from administrative work.

Fig. 8 Church Leaders’ Fears

In free response, 180 respondents named at least one worry. The clear biggest worries are outsourcing sermons, teaching, and pastoral work, laziness, over-reliance, cognitive offloading, and replacing human relationships and presence.

Church Leaders’ Hopes in Their Own Words

  • AI has the ability to increase efficiency, especially in areas like policy, procedure, team scheduling, back-end systems and structures. This is necessary stuff churches need but also takes away from time for discipleship. So the more time I can get back to disciple, pray for, and teach my congregation, the better.

    Senior pastor, non-denominational church, 41–200
  • I’m in a small, under-staffed church in an under-resourced community. I don’t want to stop doing this work, but the weight is heavy. AI helps make the load lighter in some areas.

    Senior pastor, non-denominational church, 41–200
  • The ability for small churches to use tools to approximate a staff member that they cannot afford (graphic designer, bookkeeper, policy drafting, translation, and so on).

    Senior pastor, Baptist church, 201–400
  • I am encouraged by the thoughtful engagement our leaders have when it comes to employing AI usage. There is a real desire to see AI as a tool, but not to shortcut activities that are formative for us as leaders.

    Senior pastor, Baptist church, 201–400
  • It can free up individuals to do less toil and more labor, and spend more time on things they are called to and made for.

    Executive pastor, church of 401–800
  • I am a small-church pastor. I have a part-time church administrator who is often busy with other tasks. For tasks that I would feel comfortable handing to her entirely, AI is indispensable. AI will make an exceptional administrative assistant, but anytime there is content created, I am deeply concerned.

    Senior pastor, Baptist church, 41–200
  • I’m very uncomfortable with AI overall. I don’t have a deep reason other than that it feels suspicious. However, we feel comfortable letting AI programs draft translations of the sermon manuscript for the sermon translator. It saves so much time preparing. The human volunteer translator of course meticulously reviews the AI-translated script.

    Senior pastor, non-denominational church, 41–200
  • People have found their way to our church by asking tailored questions to AI about kinds of churches that are reformed, exegetical, missional, and so on.

    Senior pastor, non-denominational church

Church Leaders’ Fears in Their Own Words

  • The ministry of the word isn’t algorithmic. It’s a Spirit-led work informed by a concrete locality in the context of human relationships. When I see people outsourcing sermons, Bible studies, and pastoral care to robots, I fear we’re bowing to the idol of machine efficiency. Ministry is, by design, inefficient. God gave us limits, and loving them is a blessing. AI sometimes offers us the same promise the serpent did, ‘You can be as gods.’

    Senior pastor, Presbyterian church
  • I’m really worried about it. First, I really hate the idea of pastors using this at all in the sermon process. The process itself shapes us, and it takes time to grab a book, search for what I’m looking at, and so on. To outsource that to AI changes the experience entirely. Additionally, I’m worried guys will simply start using AI for a lot of work they consider to be ‘busy work’: small-group questions, Sunday school outlines, and so on. We need to fight against that as much as we can.

    Senior pastor, PCA church, 41–200
  • I am extraordinarily concerned that pastors will use AI in place of the Holy Spirit. Taking shortcuts for things that are not meant to be shortcuts . . . preaching, prayer, pastoral care, and counseling.

    Senior pastor, non-denominational church
  • Increasing biblical and spiritual illiteracy. People forgetting how to write their own prayers or think. Pastors forgetting how to preach. I’m also worried it will kill community. When you used to go to other people for help, now people will go to AI.

    Senior pastor, non-denominational church
  • It’ll be really tempting to have AI write sermons and devos for our members. We’re drafting an AI policy. I’ve already had to realize where I need to pull back on some of my AI usage and research. I fear we’ll have a whole generation of new pastors who will become incredibly dependent on it and will be missing vital human skills to pastor faithfully.

    Associate pastor, Southern Baptist church
  • I have found that AI (often?) returns misleading (at best) or wrong (at worst) answers to questions. I have often checked the sources to its response and found that its answer is not even contained in the citation. When I follow up and ask it concerning this, it will say that it ‘hallucinated’ or was ‘wrong.’ My concern is that most people will not check AI’s sources, and so they will just believe whatever summary statement they give—even though it may be unfounded and totally incorrect. My other concern is with the LLMs and who is choosing what goes in them. That will naturally slant the AI bot’s response to things.

    Senior pastor, non-denominational church, 41–200
  • (1) Not allowing the Holy Spirit to influence what we are creating or diminishing our perceived need for Jesus; (2) Data protection particularly with the congregation’s personal data; (3) Training the next generation to think well or learn/develop (Are we trading development for efficiency?); (4) Even though AI is supposed to free up time and give the opportunity for more relationships and connectedness, does it actually result in us being busier and more isolated (similar to the results of social media)?

    Executive pastor, Southern Baptist church
  • 1) Creating the façade of human care and shepherding; 2) Diminishing our role as image-bearers who are called to create and cultivate; 3) Artificial content that cheapens the sweat equity invested in thinking through a sermon/devotional and thinking for the people receiving it.

    Executive pastor, Southern Baptist church
  • Cognitive decline, dishonesty, laziness, losing one’s pastoral voice, pulling up the ladder on interns/pastoral assistants (harder to find them work/justify hiring them), reducing community thought (AI as the new conversation partner).

    Associate pastor, Baptist church
  • I am concerned that people will use it as a shortcut in discipling and being discipled, which will hinder their sanctification and stunt growth. I am concerned that it will make it more difficult to accurately assess a person’s spiritual maturity based on their writing and communication. I worry that men will rely more on AI than the Holy Spirit in teaching and leading within the church. I am concerned that the efficiency of using AI will cause people to replace Christian interaction and collaboration, which will rob the church of much of the ‘organic’ discipleship that happens when people study and teach the Scriptures in community.

    I don’t enjoy feeling like the guy with a tinfoil hat; however, there is much that concerns me about AI usage in the church.

    Senior pastor, Baptist church, 201–400
  • There aren’t clear lines on what is ethically acceptable for pastor/church usage. My rule is, don’t ask it to do anything you wouldn’t ask an assistant to do for you and make sure your heart is present in any content it gives you. I have ADD and starting things is hard. I know what my heart wants to do, so I often prompt ideas for a starting point. Regardless, I am not resolved about my own guidelines, and I am sure others’ lines are even blurrier. I was just at a conference and one of the speakers prayed a prayer from ChatGPT.

    Senior pastor, non-denominational church, 41–200

Areas of Discussion for Your Leadership Team

Discussion Questions for Your Elder Board and/or Staff Team

  1. When each of you took the quiz, what persona fit you and do you agree with the one you were matched with?
  2. Of the 25 use cases, which ones would you classify as clear yes and clear no?
  3. What are your biggest fears with AI use in our local church?
  4. What are your biggest hopes with AI use in our local church?
  5. Where do we draw the line in sermon prep?
  6. Where do we draw the line in congregation or public facing communications?
  7. Are there any areas where we are having a hard time “agreeing to disagree”?
  8. What policies or action steps do we need to take in light of this conversation?

Written AI Policies

It is good for a local church to have a written AI policy. It is wise that written policies be at least understood by the leadership team, staff, and other key stakeholders, and it is worth considering updating employee handbooks to this end. We also recommend looking at the conclusion chapter on 10 Resolutions for AI Use in The AI Apocalypse book.

We have received lots of requests for examples of written church policies on AI. Many such policies can be found deploying LLMs by prompts like, “Find me ten excellent written AI church policies from gospel-centered churches in networks like The Gospel Coalition.”

What follows is a thorough list of categories that could be included in a written AI policy. These categories are far more thorough than what any church needs, and the overwhelming majority of these categories will either not apply or be too specific for your specific church’s needs. That being said, the list seeks to be comprehensive in the event you find an area of relevance that you had not thought of already.

There are 10 main areas, and nested underneath each of those main areas are subheadings worth consideration:

  1. 01. Theological & Ethical Framework
    • General AI Posture
    • Doctrine of Imago Dei and Technology
    • Creator-Creature Distinction
    • Theology of Embodiment
    • Ethics of Truthfulness
    • Doctrine of Providence of God
    • Doctrine of Common Grace
    • Theological Limits of Automation
    • Universal Ethical Principles
  2. 02. Pastoral Care, Sermons, & Sacred Functions
    • Pastoral AI Posture
    • Sermon Preparation & Exegesis
    • Sermon Drafting & Homiletics
    • Teaching, Curriculum, & Catechesis
    • Counseling, Crisis, & Soul Care
    • Prayer, Liturgy, & Worship
    • Membership, Discipline, & Benevolence
    • Pastoral Chatbots & Spiritual Q&A
    • Translation & International Ministry
    • Children, Youth, & Vulnerable Adults
  3. 03. Governance, Elders, & Leadership Oversight
    • Governance AI Posture
    • Policy Scope & Definitions
    • Governance & Human Accountability
    • Approval Authority & Decision Rights
    • AI Governance Committee
    • AI System & Use-Case Inventory
    • Risk Classification & Impact Assessment
    • Use-Case Approval & Experimental Use
    • Vendor Procurement & Due Diligence
    • Board & Elder Operations
    • Crisis, Complaints, & Appeals
    • AI Literacy & Training
  4. 04. Staffing, HR, & Internal Operations
    • Staff AI Posture
    • Approved & Prohibited AI Tools
    • Staff Acceptable Use
    • Productivity & Human Formation
    • Recruiting & Hiring
    • Performance & Employment Decisions
    • Volunteer Management & Safeguarding
    • Internal Communications & Meetings
    • Administrative Operations & Knowledge Management
    • IT, AI-Generated Code, & Accessibility
  5. 05. Public Communications & Creative Media
    • Communications AI Posture
    • External Communications & Social Media
    • Transparency & AI Disclosure
    • Graphic Design & Visual Media
    • Presentations & Data Visualizations
    • Audio, Video, & Podcasts
    • Synthetic Voices, Avatars, & Deepfakes
    • Testimonies, Quotations, & Personal Stories
    • Brand & Representational Integrity
    • Website Assistants & Public Information
    • Translation & Global Publication
  6. 06. Agentic Systems & Automated Workflows
    • Agentic AI Posture
    • AI Autonomy Levels
    • Public-Facing AI Agents
    • Internal Operational Agents
    • Identity, Credentials, & Least Privilege
    • Human Oversight & Approval
    • External Communication & Transaction Authority
    • Logging, Monitoring, & Cost Controls
    • Prompt Injection & Agent Security
    • Kill Switches, Rollback, & Human Fallback
    • Testing, Reauthorization, & Prohibited Functions
  7. 07. Data Management, Analytics, & Financial Systems
    • Data & Financial AI Posture
    • Data Governance & Quality
    • Financial Modeling & Forecasting
    • Accounting & Audit Support
    • Treasury, Payments, & Purchasing
    • Donor & Giving Analytics
    • Donor Communications & Benevolence
    • Dashboards & Predictive Analytics
    • Attendance, Membership, & Engagement Analytics
    • Grants & Fundraising Analysis
    • AI-Generated Reports
  8. 08. Data Security, Privacy, & Confidentiality
    • Privacy & Security AI Posture
    • Data Classification & Minimum Necessary Use
    • Sensitive Pastoral & Spiritual Data
    • Children, Vulnerable Adults, & Biometrics
    • Enterprise AI & Data Controls
    • Identity & Access Management
    • Retention, Deletion, & De-Identification
    • Cybersecurity & AI-Specific Threats
    • Privacy, Consent, & Individual Rights
    • International Data Transfers
    • Privacy & Security Incident Response
  9. 09. Legal, Regulatory, & Intellectual Property Compliance
    • Legal & Regulatory AI Posture
    • Legal Governance & Counsel Review
    • Federal, State, & Local AI Laws
    • Employment & Civil Rights
    • Children, Education, Biometrics, & Recording
    • Copyright, Authorship, & Training Inputs
    • Trademark, Defamation, Voice, & Likeness
    • Contracts, Terms of Service, & SaaS
    • GDPR, EU AI Act, & International Compliance
    • Nonprofit, Tax, & Fundraising Compliance
    • Accessibility & Civil Access
  10. 10. Audit, Reporting, & Lifecycle Management
    • Audit & Lifecycle AI Posture
    • Logging, Monitoring, & Transparency
    • Internal & External Audits
    • Leadership & Congregational Reporting
    • Incident Response, Complaints, & Corrections
    • Change Management & Reapproval
    • Vendor Reassessment & Renewal
    • Policy Review & Versioning
    • System Retirement & Secure Deletion
    • Enforcement & Corrective Action
    • Maturity & Continuous Improvement

Methodology

We acquired 208 responses from pastors in our TGC church network from August 21 to September 14, 2026. Respondents rated their relative comfort on a scale of 1–10 and also noted how frequently they used it on a scale from never to daily. They also had the opportunity to share qualitatively in open-ended free-response questions about what encourages them, what worries them, and the successes and failures they have seen thus far in the usage of AI in their church setting. All quotes from Part 3 have been anonymized, and there is no respondent-level data published here.

The four personas were created through k-means clustering.

There are several limitations worth noting for this survey. Conducting surveys like this one is extremely difficult to do methodologically and is essentially possible from a random sample standpoint. Even if we could do a random sample, it would not be very valuable in terms of actionable insights, as we believe that most people reading this survey primarily want to know what other gospel-centered churches are thinking and aren’t interested in what those outside that theological stream are thinking. For these reasons, we elected not to attempt a survey conducted with a random sample and peer-reviewed pathways. As a result, the survey is subject to most of the weaknesses associated with online opt-in surveys. Respondents were also given a free PDF of The AI Apocalypse book as a thank-you for participating.

The sample size is large enough to be statistically representative of the several thousand churches connected to The Gospel Coalition. Further, the data set didn’t move more than 2 percent on any question after the first 75 responses.

The data visualizations were created in Claude. All writing in this report is 100 percent human-authored and in accordance with TGC AI Guidelines for Submissions.

Figures may be shared with attribution: The Gospel Coalition, “AI Church Use: 2026 Report”