Initial draft generation. The AI produces an initial draft of the study following the five-point structural framework above. This draft is not the finished product—it is raw material that requires substantial pastoral editing. But it provides a foundation that significantly accelerates the production process. Instead of beginning with a blank page, we begin with a structured draft that already incorporates relevant cross-references, historical context, Greek and Hebrew exegesis, and application suggestions.

Quotation research. Our studies include a minimum of two quotations from Baptist or Reformed authorities for each main verse under examination. The AI assists in locating relevant quotations from John Gill, C. H. Spurgeon, John Bunyan, Benjamin Keach, Albert N. Martin, and other approved sources. Here, however, a critical caveat applies—and it applies with such force that I have made it a non-negotiable rule in our citation standards.

Citation integrity is paramount. Quotation marks are used only for verified verbatim quotes, and every such quote must include the author’s name, the title of the work, the date of publication, and the page or paragraph number. If the AI cannot verify that a quotation is verbatim—and AI systems are notorious for generating plausible-sounding but fabricated “quotations”—then the idea is attributed without quotation marks, using formulations such as “Spurgeon taught that…” or “As Gill observed in his commentary on this passage…” This rule is absolute, and violating it would undermine the credibility of every study we produce. I would rather attribute an idea loosely than attribute a fabricated quotation precisely.

Theological review. After the AI generates a draft, I review it line by line for theological accuracy, pastoral sensitivity, and stylistic consistency. I correct errors, add personal observations drawn from my own experience at Casablanca or elsewhere, adjust the tone for our specific audience, and ensure that every doctrinal statement aligns with the Second London Baptist Confession of Faith (1689). Valentín does the same. The final product is a human-authored study that has been substantially assisted—but not authored—by AI. Start to finish, we can print a high-quality study and have it ready for the prison in under 2 hours.

Theological Guardrails

AI systems are trained on vast corpora of text that include every theological tradition from high Calvinism to rank Pelagianism, and they have no inherent capacity to distinguish between truth and error. Left to their own devices, they will produce output that reflects the statistical average of their training data—which, given the overwhelming dominance of Arminian and broadly evangelical content on the internet, means they will default to theological positions that are incompatible with our confessional commitments. Our guideline document, therefore, includes explicit theological guardrails that the AI must observe:

No universal atonement language. The AI must never say “Jesus died for you” in an evangelistic context, because we do not know whether a given individual is among the elect. The approved formulation is “Jesus died for sinners”—a statement that is both theologically precise and evangelistically compelling. Every sinner who hears that Christ died for sinners is invited to flee to Him; the question of particular election is God’s business, not the evangelist’s.

No misuse of Revelation 3:20 or 2 Peter 3:9. Revelation 3:20 (“Behold, I stand at the door, and knock”) is addressed to a church, not to the unconverted, and using it as an evangelistic invitation distorts its meaning. 2 Peter 3:9 (“not willing that any should perish”) refers to God’s patience toward the elect mentioned in 1:1, not to a universal salvific will. Both passages are frequently abused in popular evangelism, and our AI must not perpetuate that abuse.

No “accept Christ” language. The synergistic framework implied by “accept Christ into your heart” or “make a decision for Jesus” is theologically backward. Christ accepts us; we do not accept Him. The sinner’s response to the Gospel is repentance and faith—both of which are themselves gifts of God (Ephesians 2:8-9; 2 Timothy 2:25)—not an autonomous decision that activates salvation. Our studies use language consistent with monergistic soteriology: God regenerates, the Spirit illuminates, Christ saves, and the sinner responds with the repentance and faith that God Himself grants.

No emphasis on personal choice apart from predestination. This does not mean that we deny human responsibility—we affirm it vigorously, as does the 1689 Confession. But responsibility is exercised within the framework of divine sovereignty, not apart from it. The AI must never present salvation as contingent solely on the individual’s decision, as though God’s eternal decree were waiting on human permission to take effect.

Preferred Theological Authorities

Our guideline document establishes a hierarchical preference for theological authorities, reflecting both our Baptist identity and our commitment to the broader Reformed tradition.

Historic Baptist sources receive the highest priority: John Gill, whose Exposition of the Old and New Testament remains unsurpassed in its exegetical thoroughness; John Bunyan, whose pastoral writings speak with unique power to prisoners; Benjamin Keach, whose allegorical and devotional works nourished generations of Particular Baptists; Roger Williams, Isaac Backus, and John Leland, whose witness for liberty of conscience defines the Baptist political heritage; Andrew Fuller and William Carey, whose theological and missionary vision launched the modern missions movement. C.H. Spurgeon’s ideas or quotations appear in nearly every study and sermon of Bautistas Históricos. John MacArthur, Walter Chantry, and Albert N. Martin also make appearances.

Waldensian and Anabaptist sources occupy a secondary but important position, reflecting the pre-Reformation heritage of believers’ baptism, congregational independence, and separation of the local church from the satanic state.

Presbyterian and Reformed sources are used extensively, too, including John Calvin, Charles Hodge, and George Whitefield. Charles Haddon Spurgeon, previously mentioned, though a Baptist, drew freely on the Presbyterian and Puritan traditions in his magisterial Treasury of David for Psalms exposition, and in his multitude of sermons.

Contemporary Reformed Baptist sources include Albert N. Martin, whose preaching and pastoral theology have profoundly shaped my ministry, and whose evening prayer format we follow in our studies.

Additional authorities include John Broadus (A Treatise on the Preparation and Delivery of Sermons), John Dagg (Manual of Theology), and A. T. Robertson (Word Pictures in the New Testament), each of whom contributes specific expertise—homiletics, systematic theology, and Greek exegesis, respectively—to our Bible study preparation.

A Concrete Example

Allow me to illustrate this process with a specific Bible study we prepared in March 2026 on Jeremiah 11:18-23, titled “¿Confías en que Dios ve tu sufrimiento y defenderá tu causa?”—“Do you trust that God sees your suffering and will defend your cause?”

The study opens with a hook designed to arrest the inmate’s attention: You are not the first person to be persecuted by those closest to you. Jeremiah—one of God’s most faithful prophets—was the target of an assassination plot organized by the men of his own hometown, Anatot, probably including his own family members. Hence, his own neighbors, and likely his own relatives, conspired to kill him because they could not tolerate his message. The connection to the prison experience is immediate and visceral: many inmates have been betrayed by family members—turned in by brothers, abandoned by wives, disowned by parents. The experience of family betrayal is among the most psychologically devastating aspects of incarceration, and Jeremiah’s story speaks directly to it.

The Contexto (Context) section establishes the historical background: Jeremiah’s ministry during the reign of Josiah and the subsequent decline of Judah; the nature of the covenant lawsuit in Jeremiah 11; and the specific circumstances of the Anatot conspiracy and of being cast into a miry cistern to die. The AI-generated draft provided accurate historical context that I then supplemented with observations from Gill’s commentary and personal pastoral application.

The study cross-references Psalm 35 (“Plead my cause, O Lord, with them that strive with me”), Psalm 109 (David’s imprecatory prayer against his betrayer), Matthew 10:36 (“And a man’s foes shall be they of his own household”), and Romans 12:19-21 (“Vengeance is mine; I will repay, saith the Lord”). These cross-references, initially suggested by the AI and verified by pastoral review, create a biblical-theological framework that connects Jeremiah’s experience to the inmate’s experience through the lens of divine justice and sovereignty.

The study incorporates John Bunyan as a specific example—a man who was imprisoned not by strangers but by a system supported by his own countrymen, a Baptist pastor who suffered for his faithfulness and who wrote from prison the most influential allegory in the English language: The Pilgrim’s Progress. Bunyan’s story resonates with prisoners because he was a prisoner, and his witness demonstrates that God’s purposes are not thwarted by incarceration but advanced through it.

The Conclusión (Conclusion) speaks separately to two audiences: the unconverted listener is warned that the same God who defended Jeremiah against his persecutors also judged the wicked men of Anatot—and that without repentance, divine judgment awaits all sinners. The believing listener is encouraged that the God who saw Jeremiah’s suffering and vindicated his cause sees their suffering also, and that His justice, though delayed, is certain. “Jehová de los ejércitos… juzga con justicia”—“the LORD of hosts… judgeth righteously” (Jeremiah 11:20).

Why This Approach Works

Several features of our AI-assisted approach account for its effectiveness:

The clear theological framework prevents doctrinal drift. Every study follows the same five-point structure, ensuring that the Gospel is always central and that the doctrines of grace are always present. The AI cannot omit the Malas Noticias section or soften the Buenas Noticias into a vague affirmation of God’s amorphous love, because the structural framework does not permit it.

Prison-specific application makes the theology concrete. Abstract doctrine, however true, does not minister to a man in a concrete cell. Our studies consistently bring the biblical text into contact with the daily realities of incarceration: isolation, betrayal, injustice, temptation, despair, and hope.

The approach treats inmates as capable of serious theological reflection while respecting their educational realities. Most of these men have not finished high school. Many have not finished grade school. We do not pretend otherwise. But we have learned—by experience, not by theory—that educational attainment and intellectual capacity are not the same thing. A man who dropped out of school at fourteen may have spent the subsequent twenty years surviving by his wits in environments that would break a university professor. He can think. He can reason. He can engage with serious ideas when they are presented clearly and concretely. We use Greek and Hebrew terms, but we transliterate them, define them, and explain why they matter. We cite historical theologians, but we introduce them—“John Bunyan, the Baptist pastor who spent over twelve years in an English prison”—so that the inmate knows who is speaking and why he should listen. We present complex doctrinal arguments, but we anchor them in the daily realities of prison life. The inmates at Casablanca have demonstrated repeatedly that they can engage with material at this level—and that they hunger for it, having been fed nothing but theological pablum by the prosperity preachers and Pentecostal entertainers who dominate the prison religious landscape. Having the AI Bible study printed allows them to review and reflect further on what the pastor shared.

Citation integrity protects credibility. Our absolute rule against unverified quotation marks means that every attribution in our studies can be trusted. In a prison environment where trust is scarce and suspicion is the default posture, this kind of integrity matters enormously.

The division of labor is sustainable. AI handles the time-intensive tasks of cross-referencing, draft generation, and quotation research. The pastors handle the tasks that require theological judgment, pastoral sensitivity, and personal knowledge of the inmates. This division allows us to produce eight or more studies per month—a pace that would be unsustainable if every study were written entirely from scratch.

The approach is operational, not experimental. We have been using AI in this capacity since mid-2025, producing studies on a regular schedule. This is not a pilot program or a proof of concept. It is an established component of our ministry workflow.

Implications for the Church

The implications of this approach extend far beyond our ministry at Casablanca. Any church with access to AI—which is to say, effectively, any church with access to the internet—can now prepare quality Bible study materials for prison ministry. This technology democratizes what previously required a seminary-trained pastor working full-time on curriculum development.

Consider the small rural church that wants to begin a prison ministry but lacks the pastoral resources to prepare weekly or monthly Bible studies. With a well-crafted set of theological guidelines (which we make freely available at bautistashistoricos.com), that church can use AI to generate initial drafts that a lay elder or deacon—someone with sound doctrine but without formal homiletical training—can review, correct, and finalize. The quality of the output will be directly proportional to the quality of the theological guidelines, but even a modest set of guardrails will produce material that is superior to the theologically vacuous handouts that characterize much of contemporary prison ministry.

Consider the missionary in a foreign country who needs to produce Bible study materials in a language he is still mastering. AI can generate accurate text in virtually any major language, and a native speaker can then review and refine it. This is precisely how Valentín and I operate: I provide the theological content and structural framework in English, the AI generates output in Spanish, and Valentín verifies the linguistic and cultural accuracy if there is any doubt. Usually, my Spanish is more than sufficient to read and detect errors alongside Valentín, but we work hand in hand to bring the word of God to prison disciples and seekers in the most efficient and effective way possible in a tent-making ministry. The result is a Spanish-language Bible study that reflects Reformed and Baptist theology with the precision of a trained theologian and the fluency of a native speaker.

Consider the chaplain managing multiple units in a large prison, responsible for hundreds or thousands of inmates across varying levels of literacy, spiritual maturity, and doctrinal background. AI can help produce differentiated materials—simpler studies for new converts, more advanced studies for mature believers, evangelistic studies for the unconverted—at a pace that no individual could match working alone.

The technology is here. The question is not whether the church will use it, but whether the church will use it wisely—with theological guardrails, pastoral oversight, and a clear understanding that AI is a tool in the hands of a servant, not a replacement for the servant himself. “La preparación del corazón es del hombre; mas la respuesta de la lengua es de Jehová”—“The preparations of the heart in man, and the answer of the tongue, is from the LORD” (Proverbs 16:1). The AI prepares; the Lord answers. The pastor’s task is to ensure that what the AI prepares is worthy of the Lord’s use.

Action Steps

1. Develop a theological guideline document for your ministry—or start with ours. Our two-page prison Bible study template and eighteen-page sermon preparation guide are available for download at bautistashistoricos.com. Download them, study their structure, and modify them in line with your confessional commitments and ministry context. If you prefer to build from scratch, establish clear doctrinal parameters before using AI: what confession of faith governs your ministry? What theological errors must the AI avoid? What structural framework ensures consistency? Even a twenty-page summary of your essential convictions will dramatically improve the quality of AI-generated output. Upload them to Claude or another AI service and prompt it to make a study based on your outline.

2. Begin using AI as a research assistant before using it as a drafting tool. Ask it to find cross-references for a passage you are studying. Ask it to summarize the historical context of a biblical book. Ask it to identify relevant quotations from trusted theologians. Verify every output against your own knowledge and reliable sources. This will teach you both the technology’s capabilities and its limitations.