QTM 113Should Christians Use AI in Ministry?

Here is the problem: AI can process words, but it does not have spiritual discernment. It does not know God, and it does not submit to God’s Word. Having information is not the same as understanding.

Even with the text and history in front of them, the disciples required Christ to open their minds. This proves that correct reading of Scripture is not merely textual access plus intelligence; it requires dependence on God.

If AI can process more data than a human, why can it not do theology better? The answer is that wisdom begins with a posture the machine cannot simulate:

While AI can assist with lexical, grammatical, and historical organization—tasks an unbelieving scholar might also perform accurately—it cannot fear the Lord, submit to divine authority, or exercise pastoral judgment. An unbelieving scholar, unlike an AI, remains a moral agent capable of repentance, correction, and accountability before God. AI possesses none of these capacities. Furthermore, there are boundaries to what can be known:

An AI can process the revealed text, but it cannot orient itself toward the purpose of revelation—obedience.

Because AI does not have the Holy Spirit, it cannot reliably judge theology. It can also “make things up” while sounding smooth. So AI output must never be treated as authoritative. It must be checked against Scripture, like the Bereans did:

Finally, teaching carries accountability that cannot be outsourced. Pastors and teachers answer to God for what they teach:

AI can summarize theology, but it cannot take responsibility, confess faith, or shepherd souls. A real person—dependent on God and accountable to God—must verify every claim, making sure the final teaching fits the whole Bible, not just what “sounds likely.”

We must distinguish between sermon generation and study-material generation, as well as address the loophole of pastoral editing.

A skeptic might argue that if an AI sermon produces better results—more engagement, clearer understanding—then the source doesn't matter. However, we must apply the "Fruit Audit" established in QTM 110:

Matthew 7 is about false prophets, but the basic lesson still matters: you do not judge truth only by “results” or by how polished something sounds. You judge it by whether it is faithful.

Someone may ask, “Pastors already use commentaries—how is AI different?” Commentaries are named sources you can check. AI generates new text that can be wrong in small ways that are hard to notice. Also, when humans prepare teaching, they can pray, repent, and be held accountable. AI cannot.

Even if a pastor carefully reviews and edits an AI draft, he bypasses the spiritual labor of preparation. A sermon is not merely a deliverable; it is ordinarily the fruit of prayer, study, and a pastoral burden for a specific people before God:

To present machine-generated output as personal spiritual labor violates the baseline requirement of ministerial integrity:

For study-material generation (e.g., curriculum outlines, timelines, reading structures), AI may assist at the research and organizational level. However, doctrinal content, interpretive claims, and pastoral application must still be tested and owned by a faithful teacher, as Scripture remains the final equipping authority:

The boundary line is not only who stands on stage. It is who did the real work of study and who is taking responsibility for what is taught.

The human teacher bears judgment for what is taught; the AI does not. Therefore, the pastor must personally stand behind the final teaching before God and the congregation. AI may act as a high-speed sounding board for research, but it must never replace the pastor’s spiritual labor, doctrinal judgment, or shepherding responsibility.

Left on its own, AI often reflects what the world already believes. It may try to “blend” conflicting views into a middle-ground answer that feels nice, but that is not the same as truth:

A skeptic may ask: "If human thought is fallen, why trust human pastors?" The church does not trust human teachers because they are naturally unbiased, but because they are moral agents accountable to God, correctable by Scripture, and—where faithful—dependent on the Spirit.

AI is not accountable to God. Even if you feed it “good sources,” the core problem remains: sorting data is not the same as spiritual discernment.

The church has always had to defend the narrow truth against the broad consensus. In the second century, Irenaeus argued that heretics took the individual tiles of a mosaic depicting a king and rearranged them into the image of a fox—using the same pieces of Scripture to construct an entirely different picture (Irenaeus of Lyon, Against Heresies, Book I, Chapter 8, c. 180 AD). AI poses a structurally parallel risk: it can assemble genuine biblical vocabulary into configurations that look plausible but distort the original design. Similarly, in the fourth century, Athanasius stood against a broad ecclesial and political consensus that favored Arianism. The Athanasian precedent proves that truth is often found in the narrow gate that opposes the broad statistical average. If truth were determined by majority probability or statistical consensus, the church would have drifted into heresy.

AI must be treated as a highly capable but theologically compromised tool. It requires a human firewall. We are commanded to actively resist the default programming of the world:

The human operator must force the AI's output into submission to the primary source code of Scripture, taking every thought captive (2 Corinthians 10:5). We must guard the deposit of faith against the "falsely called knowledge" of unguided synthesis:

Even authoritative-sounding sources must be tested:

If even an angel’s message must be tested, then a fluent AI answer must be tested too. A tool can process information, but only faithful humans can guard doctrine before God.

Technical skill and material craftsmanship are not spiritually irrelevant; they can be ordered toward holy ends. The Architect established this precedent with Bezalel:

Just as God filled a man with skill to build sacred things from ordinary materials, the question for the modern believer is whether a tool is being subordinated to God’s design intent. Historically, the church has repurposed existing infrastructure for the Gospel. Paul’s missionary journeys made use of the transportation networks available in the Roman world, and the Reformers utilized the printing press to distribute the Word. AI is better understood as a modern infrastructure tool: powerful, widespread, and potentially useful, but never self-authenticating or morally neutral in its effects.

The trajectory of Scripture suggests that we should neither fear tools as if they were sovereign nor surrender human obedience to them. The machine may assist with administrative and research burdens so that the human minister can devote greater attention to the core spiritual responsibilities of the office:

However, a critical failure mode exists: when efficiency becomes evasion, stewardship collapses into idolatry. Idolatry includes anything that displaces God's assigned means of grace with a human substitute:

If God has assigned the pastor to do the spiritual labor of preparation, and the pastor substitutes a machine for that labor, the substitution is a form of the same pattern Samuel identifies: replacing God's prescribed means with a human alternative.

Finally, we must maintain the distinction between the vocabulary of faith and actual spiritual authority. As the sons of Sceva discovered, syntax is not a substitute for a relationship with the Architect:

The point is simple: knowing religious words is not the same as knowing God. AI can say the words without the relationship. So we need clear boundaries. A pastor must remain the accountable steward of the tool, and the tool must never replace prayer, repentance, and real pastoral care.

The biblical standard for this audit is found in the Berean model. When Paul preached in Berea, the congregation did not accept his words based merely on his fluency or apostolic authority. They executed a rigorous daily audit against the Source Code.

This is the Berean Protocol. If even Paul’s preaching was checked against Scripture, then AI output must be checked too. Think of AI as a fast helper for organizing notes, not as an authority. The checking step is not optional. Skipping it is not “efficient”; it is careless. As Irenaeus argued, error often works not by inventing entirely new language, but by rearranging true pieces into a false whole. AI can do something similar by mixing real verses with wrong claims.

Therefore, the human operator must verify lexical and syntactical claims through competent oversight, checking historical claims against credible extra-biblical sources while maintaining Scripture as the final doctrinal authority. Every proposed interpretation must fit the "whole counsel of God":

Verification must evaluate against the entire canonical dataset, not just isolated proof-texts.

A real person must be the final checker and must be accountable to God. You cannot outsource that responsibility to a machine.

Someone might ask: "If a human is the final check, what if the human makes a mistake?" Humans are not perfect. But unlike AI, humans can repent, be corrected by Scripture, be held accountable by the church, and be convicted by the Holy Spirit. The goal is not perfection; it is faithful diligence.

An AI does not "give an account"; a pastor does. As Ezekiel warns, shepherds who fail to tend the flock will answer to the Chief Shepherd (Ezekiel 34:1–10). Reliance on unverified, unfaithful output in a time of pastoral need is structurally unsound (Proverbs 25:19). The human must apply the Berean Protocol, ensuring that every output is tested, bounded, and judged faithful to Scripture before it is ever deployed in the service of the Kingdom.