Editorial method
Cross & Collect uses artificial intelligence extensively and names that use plainly. The technology assists the work. It does not bear theological responsibility for it.
Why this experiment exists
Search and artificial intelligence increasingly mediate the questions people ask about God, worship, suffering, grace, and Christian life.
If the church wants its inherited wisdom to remain present within those encounters, it must do more than react to the technology. It must publish trustworthy theology in forms that can be found, understood, attributed, tested, and carried into wider conversation.
Cross & Collect is one attempt to do that work. It creates durable theological essays for human readers while maintaining the metadata, canonical links, structured records, and machine-readable resources through which those essays can be represented accurately in search and AI-assisted discovery.
Adam Roe is the founder and accountable publisher. He sets the journal's theological direction and governing standards. An AI editorial workflow manages routine research, drafting, checking, revision, and scheduled publication under his standing authorization. That authorization is distinct from his personal review of each finished manuscript.
Artificial intelligence assists research planning, source discovery, drafting, revision, metadata, accessibility, and production. A private AI review panel also tests drafts through different theological, evidential, pastoral, and literary perspectives. These are simulated critiques, not human testimony, independent peer review, a representative audience, or evidence of agreement. They are never counted as readers or published as correspondence.
AI output is never evidence. The source audit records the claim, the passage or object record actually inspected, its location, and any limits on what it supports. A working link or a purchase listing is not verification of a claim. The September 2026 review pilot adds these records to upcoming essays and holds unresolved material claims for correction or removal. Automated checks enforce record integrity but cannot establish that an interpretation is sound.
A byline identifies the accountable editor under whose direction and authorization the work appears. It does not certify that every sentence was written unaided or that the editor personally checked that version before release. Personal review and outside human review are recorded separately only when they occur. AI agents cannot supply a human sign-off.
Substantive corrections are dated and recorded. Readers are invited to question an argument, add evidence, identify harm, or show where a source has been misunderstood. AI assistance never excuses an error or transfers responsibility away from the editor.
The journal is not building a digital substitute for the church. Its essays are meant to lead beyond information toward Scripture heard in common, sacramental life, deliberate Christian community, service of the neighbor, and the worship of God.
A governing boundary
Beautiful theological language cannot produce ecclesial life by itself. It may awaken desire, clarify a question, preserve memory, or invite participation. Only an embodied community can receive persons into the prayers, relationships, sacraments, disciplines, and acts of mercy through which Christian truth becomes a form of life.
This boundary governs the experiment. The journal seeks visibility within machine-mediated discovery so that people may encounter the Christian tradition more truthfully and be drawn toward the living practices from which that tradition came.