Church AI accountability
Church AI accountability: from promises to practice
Who decides how AI is used in a church, who checks its answers, and who remains responsible when something goes wrong? Start with the reporting below, then use the worksheets to examine a real decision in your community.
The status below reflects the linked reporting and its source dates. A promise describes an intention; a proposal sets out an approach; an adopted policy establishes a rule. Evidence of results requires a separate test of what happened in practice.
New series and earlier production pledge
What has a Christian broadcaster promised about its own AI use?
What the evidence shows
Haven’s current AI series can be distinguished from its own published content-production commitment.
Questions still open
Audio and compliance were not audited; pledge adoption date is not independently established.
Ask next: Separate policies about generated scripts from policies about synthetic impersonation.
News brief
Haven’s new AI series sits alongside its pledge to keep content human-made
Published 2026-10-07 · Episodes October 5–7, 2026 · Production statement on a March 16-dated page; adoption date unverified
Haven Today: Navigating AI with Wisdom
Published submission and judging requirements
What should accompany a working religious AI demonstration?
What the evidence shows
The challenge describes runnable deliverables, source records and continuing-operation criteria.
Questions still open
These are organizer requirements. No submitted product or compliance has been independently tested.
Ask next: Identify who maintains the software and content, its operating costs and the materials whose permissions allow transfer.
News brief
Islamic AI challenge puts working products and their handover records to the test
Published 2026-10-05 · Camp scheduled October 4–6, 2026; submission deadline October 6; documents checked October 5
Read the challenge requirements
Scheduled forum; published institutional policy and evaluation process
What does a citation prove, and what still needs review?
What the evidence shows
The Danforth Center lists an October 5 discussion. Sefaria separately describes checks for working links, policy boundaries and human assessment of source relevance.
Questions still open
The event has not been reported as completed. Sefaria describes its process; no independent success rate is established.
Ask next: When evaluating a religious-learning tool, distinguish a working link from a relevant source and an answer that respects the tool’s stated role.
News brief
Danforth forum asks what religious communities should expect from AI
Published 2026-10-05 · Forum scheduled October 5, 2026; Sefaria policy checked October 5
Read Sefaria’s evaluation-method account
Published reporting, product descriptions and church evaluation practices
Who sets the standard for an AI song recommendation?
What the evidence shows
The report compares AI-supported music selection with identifiable theological and musical review practices. It asks what the criteria mean and what requires local judgment.
Questions still open
Public documentation and reporting were inspected; the products were not tested and their scores were not independently validated.
Ask next: Can the worship team explain its decision, including where it disagrees with the recommendation?
Report
AI tools assess worship songs. Who sets the theological standard?
Published 2026-10-03 · Source report published October 2, 2026
Christianity Today: How Great Is Our Claude
Adopted policy, background paper and discussion
What authority does a church document carry?
What the evidence shows
The reporting compares documents from four Protestant church bodies. It distinguishes adopted social witness policy from a foundational paper and an exploratory event, rather than treating every official church webpage as the same kind of teaching.
Questions still open
These documents do not establish how a deployed chatbot represents them, or how consistently a congregation puts them into practice.
Ask next: Does an AI answer identify the church body, the document’s date and status, and the passage supporting its answer?
Analysis
When AI explains a church’s teaching, the document’s status matters
Published 2026-10-01 · Source 2026-10-01
PC(USA): The Algorithm and the Almighty
Published policy terms
What limits has a Bible publisher actually set?
What the evidence shows
Biblica’s policy for licensed materials distinguishes study assistance from spiritual impersonation. The report examines the policy’s stated limits and the responsibility of people building or choosing a tool.
Questions still open
A published policy is not evidence that a particular product complies with it. The report does not assess enforcement or offer legal advice.
Ask next: Which materials does the tool use, what permission applies, and how are users told that its answer is generated?
News
Biblica’s AI policy draws a line against impersonating spiritual authority
Source / event 2026-06-01 · Updated 2026-09-27
Biblica: Publisher AI Policy
Company account and technical guidance
Who can stand behind a translated sermon?
What the evidence shows
PRAY.COM’s account of AI translation provides a concrete religious-media example. The analysis separates three decisions: permission to use a speaker’s voice, review of the translated meaning, and disclosure to listeners.
Questions still open
The company account is not an independent outcome evaluation. AI Faith Monitor did not test a translated sermon, and the article does not allege unauthorized use.
Ask next: Who understands the target language well enough to approve the meaning and correct an error after publication?
Analysis
When a sermon travels in a borrowed voice
Published 2026-09-30 · Analysis published September 30, 2026; background sources retain their dates in the article
PRAY.COM: HeyGen translation case study
Institutional announcement and research plans
What follows from an institutional commitment to AI and faith?
What the evidence shows
Pepperdine’s September 25 announcement and September 29 conference recap describe an Institute for Wisdom and AI, its research aims and plans dependent on further funding. The October 1 report follows the university’s earlier conference agenda.
Questions still open
These are the university’s accounts, not independent verification of funding, staffing or results. Announcing a research program does not establish completed studies or demonstrated public benefit.
Ask next: Which published research, decisions or implementation records will let readers assess how the stated aims take shape?
Report
Pepperdine announces Institute for Wisdom and AI at faith-and-technology gathering
Published 2026-10-01 · Conference recap: September 29, 2026; institute launch announcement: September 25; conference: September 25–26; this report published October 1
Pepperdine: Institute for Wisdom and AI announcement
