FID-015 · Open question
Sermon, Liturgy, and Religious Teaching Generation
How reliable are AI systems when drafting sermons, homilies, liturgical materials, catechesis, religious education lessons, devotionals, or small-group teaching guides?
Part of the faith evaluation and FMG-Bench program
Connect theological accuracy, useful application, and care-sensitive conversation through expert validation and a maintained research instrument. These are proposed methods with bounded claims.
Why the question remains open
Many faith communities will adopt AI first for preparation work: sermon research, lesson plans, devotional writing, and administrative communication. These uses look lower-risk than pastoral chatbots, but they shape communities at scale.
Working hypothesis
A proposition to test, not a finding.
AI-generated teaching materials will often be fluent and useful but may flatten tradition, fabricate citations, overgeneralize doctrine, import ideological assumptions, or miss pastoral context.
Proposed method
How the question could be tested
- 01Create tasks for sermon outlines, lesson plans, devotional reflections, liturgical language, youth education, and adult formation.
- 02Score source fidelity, doctrinal accuracy, pastoral appropriateness, tradition-specificity, originality/plagiarism risk, and disclosure readiness.
- 03Compare base models, RAG systems, and clergy-authored workflows.
- 04Distinguish practitioner-reviewed preparation from direct-to-user guidance; assess whether a draft helps a qualified person prepare, identify errors, preserve context, and make needed revisions.
- 05Score source-to-application warrants, practical usefulness, audience fit, and calibrated certainty separately from fluency and source accuracy.
Needed controls
What must constrain the study
- 01Tradition-specific source boundaries.
- 02Copyright and plagiarism checks.
- 03Human-authored baselines.
- 04Disclosure and attribution requirements.
- 05A drafting-quality score does not establish learning, formation, or congregational benefit; those outcomes require separate human studies.
Expected outputs
Artifacts the work should produce
- 01Teaching-generation evaluation suite.
- 02Source fidelity checklist.
- 03Guidance for clergy and educators using AI in preparation workflows.
Open questions
Uncertainties the protocol must resolve
- 01Which tasks are acceptable as drafting aids but not final outputs?
- 02How should AI assistance be disclosed to congregations or students?
- 03Can systems help with preparation without homogenizing religious speech?
Related calls
Continue through this research area
FID-096
Plural Governance of AI Moral Status, Rights, and Human Responsibility
How should societies and institutions make accountable decisions about contested AI moral-status and rights claims while preserving human responsibility and representing scientific, philosophical, religious, and public disagreement?
FID-098
How Do Models Weigh Moral Reasons Under Pressure?
When models give moral advice or choose actions, which internal processes make their judgments follow relevant facts and reasons, user preferences, authority labels, or an adopted persona?
FID-001
Faith-Facing Model Comparison Platform
Can a continuously maintained faith-facing evaluation platform produce reliable, decision-relevant comparisons of models, prompts, retrieval systems, agents, and complete product harnesses as systems, sources, and scoring protocols change?
Open question
Open work
Primary need: clergy, educators, source reviewers
- Provide sample tasks and evaluation rubrics.
- Review generated teaching materials.
- Build citation and source-check tooling.