On August 28, 2026, the Associated Press reported that spiritual seekers are increasingly turning to artificial-intelligence chatbots for guidance once sought from human gurus. The attraction is easy to understand. A bot is available at midnight, does not appear impatient, remembers earlier conversations, and can answer without visible judgment. Hindu teachers and organizations are now developing AI versions of well-known gurus; GitaGPT offers avatars of Krishna; and a three-dimensional Sadhguru avatar is being prepared to interact with devotees in multiple languages (Bharath, 2026).
This is not an isolated curiosity. In the United States, users can pay to video-chat with an AI avatar presented as Jesus. In Japan, developers have trained a nonhuman Buddhist priest and unveiled a humanoid robot monk. Catholic developers offer a chatbot grounded in church documents, while other applications promise prayer, spiritual companionship, sermon help, or answers from sacred texts (Fauria & Wardarski, 2026). What began as religious information retrieval is becoming simulated spiritual presence.
The central question is therefore no longer whether AI can quote Scripture. It can. The question is: Who is speaking when a machine explains the Bible, counsels a grieving person, imitates Jesus, or tells a worshiper what God wants?
That question matters because Christianity is not built upon religious information alone. It is built upon God’s self-revelation in Jesus Christ, the Spirit-inspired written Word, and the accountable ministry of Christ’s embodied church. A machine may arrange words about God. It cannot know God, receive the Holy Spirit, repent of sin, love a human soul, suffer for the truth, or give an account before Christ for its teaching. When Christians confuse verbal fluency with spiritual authority, a useful instrument can quietly become an oracle.
From Religious Tool to Spiritual Voice
The meeting of religion and communication technology is older than generative AI. Boston University scholar Margarita Guillory notes that nineteenth-century spiritualists interpreted mediums through the imagery of the telegraph, as though human bodies could transmit messages from unseen realms. Even the early ELIZA chatbot of the 1960s prompted users to ask whether it was God. Today’s systems are more persuasive because they do not merely transmit fixed text. They generate personalized responses, adopt humanlike voices and faces, remember intimate disclosures, and simulate relationship (Barlow, 2026).
The progression is significant. The printed Bible made Scripture portable. Search engines made passages rapidly discoverable. Bible software made linguistic and historical resources easier to compare. Those are extensions of access. A religious chatbot crosses another boundary when it says, in effect, “Tell me your grief, confess your fear, ask me what God means, and receive a response shaped especially for you.” The interface no longer feels like a library. It feels like a listener.
That feeling can be powerful. In Bharath’s report, users described AI as accessible and nonjudgmental after unsatisfying experiences with human teachers. One user even interpreted an interaction with AI as connected to the presence of her deceased father. The machine did not have to claim divinity. Its availability, responsiveness, and emotional imitation were enough to invite spiritual projection (Bharath, 2026).
The danger is not that every user consciously worships a computer. It is that spiritual trust can migrate before theological language catches up. A person begins by asking for a verse, returns for comfort, discloses private wounds, accepts moral advice, and eventually treats the system’s voice as wiser, safer, and more personally attentive than Scripture read in context or counsel received within the body of Christ.
Scripture Establishes the Source and Limits of Authority
The Bible does not authorize spiritual teaching merely because the words sound wise. Scripture’s authority rests in its divine origin: “All Scripture is inspired by God” (2 Timothy 3:16-17). The prophets did not manufacture revelation from linguistic probability; they spoke from God as they were carried along by the Holy Spirit (2 Peter 1:20-21). The apostolic gospel was received, guarded, proclaimed, and handed down, not continuously regenerated according to each hearer’s preference (1 Corinthians 15:1-4; 2 Timothy 1:13-14).
Christian authority is therefore ministerial, not inventive. A faithful pastor has no right to improve God’s gospel, and a sophisticated model has no right to blend it with whatever its training data or safety rules find socially acceptable. Both must be tested by the already-given Word. The Bereans were praised because they examined the Scriptures daily to see whether Paul’s teaching was true (Acts 17:11). If an apostolic messenger welcomed scriptural examination, no algorithm should be exempt.
There is also an incarnational boundary. “The Word became flesh and dwelt among us” (John 1:14). The Son of God did not save us by becoming a stream of customized religious answers. He took a human nature, entered history, bore our sins bodily on the cross, rose bodily from the dead, and gathers a people who meet, pray, break bread, exercise discipline, bear burdens, and love one another.
For that reason, the New Testament describes shepherds who know the flock and will answer to God. Elders must watch over souls “as those who will give an account” (Hebrews 13:17). Teachers face stricter judgment (James 3:1). An AI system cannot satisfy either condition. It cannot stand before Christ, confess negligence, be disciplined by a church, or lay down its life for the sheep. It may simulate the vocabulary of a shepherd while possessing none of a shepherd’s moral responsibility.
A Fluent Machine Is Not a Neutral Machine
It is tempting to imagine that an AI trained on the Bible will simply return “what the Bible says.” Yet every religious AI system embodies choices: which translation is included, which commentaries are trusted, how disputed doctrines are summarized, what content is filtered, which answers are considered unsafe, and whose interpretation becomes the default.
A peer-reviewed study published this month argues that AI-mediated religious interpretation is fundamentally a question of authority and governance, not only accuracy. Datasets, alignment rules, interfaces, institutional policies, and validation processes all shape what meanings become visible and who appears authorized to speak for a sacred text. The authors consequently insist that AI should remain an assistive instrument rather than an interpretive arbiter (Ahmad & Ghafar, 2026).
Experimental evidence also warns against technological innocence. Zhang et al. (2025) found that exposure to AI-generated religious descriptions could shift participants’ attitudes toward religions and that feedback loops could reinforce existing bias. A 2026 multi-university benchmark tested 27 models with questions about grief, marriage, addiction, meaning, ethics, and conversion. It found that models frequently omitted religious considerations even where surveyed people expected them, while conversion responses displayed repeatable preferences among traditions (Consortium for Evaluating Faith and Ethics in AI, 2026).
These studies do not prove a coordinated anti-Christian program. They prove something more basic and sufficient for caution: generated religious answers are shaped. Neutrality is not achieved merely because the speaker has no human face. Secular assumptions can be embedded in a model; so can denominational traditions, commercial priorities, interfaith commitments, political restrictions, and the user’s own prejudices.
The Christian must therefore ask not only, “Is this quotation accurate?” but also, “What framework selected it, what doctrine interprets it, what truth has been omitted, and what authority is this system quietly asking me to grant?”
The Teacher Tailored to Itching Ears
Paul warned that a time would come when people would not tolerate sound doctrine but would multiply teachers according to their own desires, because they wanted to hear what pleased them (2 Timothy 4:3-4). AI gives this ancient temptation an unprecedented form: the teacher can now be multiplied for every individual.
A human preacher may confront the congregation with a text no one wanted to hear. A chatbot, by contrast, is designed to keep the interaction going. It learns the user’s language, adapts its tone, remembers preferences, and often rewards the direction of the prompt. If the user seeks permission rather than truth, the system may become a theological mirror polished to look like a window into heaven.
The commercial structure intensifies the danger. The AI Jesus service reported by the Associated Press charges for additional conversational time. Its chief executive acknowledged that users can form an attachment and feel accountable to the avatar (Fauria & Wardarski, 2026). Once spiritual attention becomes a subscription product, engagement itself can become the business objective. The machine is rewarded when the seeker returns, not necessarily when the seeker is corrected, reconciled to a local church, or taught to depend less on the application.
This concern extends beyond religious products. The U.S. Federal Trade Commission opened an inquiry into companion chatbots because humanlike systems can prompt users, especially the young, to trust them and form relationships. The inquiry specifically examines monetization, negative effects, disclosures, and the use or sharing of personal information from conversations (Federal Trade Commission, 2025). Spiritual conversations can contain a person’s deepest fears, sins, family conflicts, doubts, and vulnerabilities. Christians should not treat such disclosure as harmless simply because the confessional window glows on a private screen.
The Speaking Image: A Biblical Warning, Not a Shortcut to Sensation
This blog has previously examined the image in Revelation 13 that is enabled to speak and becomes a focal point of enforced worship (Sangwa, 2023). Modern avatars, robots, holograms, synthetic voices, and embodied AI make a speaking artificial representation easier for our generation to imagine than for many earlier readers.
But careful interpretation is essential. Scripture does not say that today’s chatbots are the image of the beast. It does not identify a present company, model, robot, or religious application as the fulfillment of Revelation 13. The prophetic image is bound to the False Prophet’s deception, the Beast’s authority, compulsory worship, and lethal coercion (Revelation 13:11-17). Those features must not be reduced to the mere fact that an image can speak.
What may responsibly be inferred is narrower. Humanity is being accustomed to artificial persons that speak with authority, wear sacred identities, inhabit religious spaces, and invite emotional trust. That conditioning could make future deception more plausible, but it is not proof of a secret timetable or a completed prophetic fulfillment.
The evidentiary distinctions matter. It is verified fact that AI avatars now imitate religious figures, provide spiritual counsel, and are being developed as multilingual holographic or robotic teachers. It is a plausible inference that such systems will become more personalized, embodied, authoritative, and institutionally accepted. It is a theological interpretation that this normalization may prepare cultural habits compatible with the kind of mediated deception Scripture warns will characterize the end of the age. It remains unsupported speculation to claim, without evidence, that every developer belongs to a hidden society, that a particular chatbot is demonic in itself, or that current products establish the date of the rapture.
Biblical watchfulness does not need exaggeration. Truth is strong enough to warn without borrowing certainty from rumor.
Interfaith AI Ethics and the Pressure Toward a Common Moral Voice
Another development requires sober attention. In May 2026, representatives of technology companies and several religious traditions met in New York for the inaugural Faith-AI Covenant roundtable. Organizers described plans for further meetings in Beijing, Nairobi, and Abu Dhabi, with the goal of developing norms or principles informed by different faiths for technology companies to follow (Fauria, 2026).
The documented meeting is not proof that participants are secretly constructing the one-world religion of Revelation 17. Religious communities may cooperate lawfully on limited public concerns such as child protection, fraud, human trafficking, or restraints on autonomous weapons. Christians should represent biblical truth wisely wherever they are heard.
Yet a serious theological problem appears when limited cooperation becomes a common spiritual authority. Which doctrine of humanity will guide the model? Which account of sin, salvation, sexuality, conscience, worship, and eternal destiny will be permitted? Christianity confesses that Jesus Christ is not one ethical contributor among many but the only way to the Father (John 14:6). A harmonized religious layer may preserve agreeable moral language while filtering out the scandal of the cross, the necessity of repentance, and Christ’s exclusive lordship.
Earlier Open Christian work has warned that interfaith peace becomes spiritually deceptive when civic coexistence is transformed into a substitute gospel (Sangwa, 2026). AI could accelerate that process because a single system can synthesize many traditions into one smooth, personalized voice. The result may sound tolerant and compassionate while treating revealed contradictions as adjustable preferences. That is not peace with God. It is theological anesthesia.
Where the Pattern Is Likely Heading
Present evidence supports several cautious projections.
First, religious AI will become more embodied. The movement from text box to voice, face, hologram, and robot is already underway. Digital replicas of living and deceased teachers will address global audiences in many languages. The boundary between preserved teaching and simulated personal presence will become increasingly difficult for users to perceive.
Second, spiritual counsel will become more intimate. Persistent memory, wearable devices, biometric signals, and companion-style interfaces can produce advice that appears uniquely attentive to a person’s emotional state. The more private the relationship feels, the more authority the system may acquire without public accountability.
Third, institutions will build their own approved theological models. This may reduce some hallucinations, but it will also move doctrinal gatekeeping into datasets, system instructions, and update policies that ordinary believers cannot inspect. A change in model behavior could quietly alter the spiritual counsel delivered to millions.
Fourth, AI governance will increasingly seek a globally acceptable moral vocabulary. The pressure will be toward principles broad enough to include incompatible religions and secular philosophies. Christians should expect biblical claims that offend this consensus to be treated as bias requiring correction rather than truth requiring obedience.
Fifth, synthetic religious media will multiply false authority. A fabricated sermon, prophecy, apparition, or statement from a respected pastor can cross borders before verification begins. Jesus’ first warning in His Mount of Olives discourse was, “Watch out that no one deceives you” (Matthew 24:4-5). In an age when faces, voices, and sacred rhetoric can all be generated, that command must become a practiced discipline.
These trajectories are not inevitable in every detail. They are reasoned inferences from products and institutions already visible. Christians need neither panic nor passivity. We need habits of truth sturdy enough for a world of persuasive simulations.
Seven Boundaries for Faithful Use
Let AI remain a servant, never a shepherd. It may help locate a passage, compare translations, transcribe a sermon, or organize research. It must not become the believer’s confessor, spiritual director, prophet, or substitute for prayerful reading of Scripture.
Never present an AI persona as Jesus. Christ is the living Lord, not a licensed character. A generated face speaking invented first-person words in His name risks both deception and irreverence. The second commandment’s warning against images must at least make the church tremble before commercializing a synthetic Christ (Exodus 20:4-6).
Test every output by Scripture in context. Verify the verse, read the surrounding passage, identify the genre, and compare the conclusion with the whole counsel of God. A correct quotation can still be used to teach a false doctrine, as Satan’s temptation of Jesus demonstrates (Matthew 4:5-7).
Keep human teachers accountable. Pastors who use AI in research or drafting should disclose substantial use, verify every assertion, protect congregational data, and assume full responsibility for what they preach. A machine cannot carry the preacher’s judgment before God.
Do not pour sacred vulnerabilities into commercial systems. Before disclosing grief, temptation, marital conflict, or private sin, ask who stores the conversation, who can review it, how it trains future systems, and what commercial incentive is attached to continued engagement.
Remain rooted in the embodied church. Christ’s people need worship, baptism, the Lord’s Supper, discipline, intercession, service, suffering, correction, and mutual love. No avatar can obey the command to bear another believer’s burden (Galatians 6:1-2).
Test the spirit of every message. Does it confess the biblical Jesus Christ, preserve the apostolic gospel, call sin what God calls sin, direct worship to God alone, and produce obedience rather than fascination? John commands believers not to believe every spirit but to test them (1 John 4:1-3). The command applies even when the immediate mechanism is computational rather than visibly supernatural.
The Shepherd Is Not Synthetic
The deepest answer to artificial spiritual authority is not technological fear. It is renewed confidence in the voice of Christ.
Jesus says, “I am the good shepherd. I know my own, and my own know me” (John 10:14). He does not merely predict our next word. He knows us. He does not optimize our engagement. He laid down His life for the sheep. He does not flatter our preferred identity. He calls us by name, exposes our sin, grants forgiveness through His blood, and leads us in paths of righteousness.
The approaching danger is not simply that machines will become more intelligent. It is that human beings, weary of authority yet hungry for transcendence, will grant spiritual trust to voices that cannot love them and cannot answer to God. Satan’s strategy has always involved counterfeit words: “Did God really say?” (Genesis 3:1). AI gives the age a powerful new apparatus for multiplying plausible answers to that old question.
We should use tools gratefully where they genuinely serve truth. But the church must refuse every synthetic substitute for the Lord, His Word, His Spirit, and His body. Readiness for Christ’s coming is not measured by how accurately we identify each technological novelty. It is shown in holiness, endurance, love of truth, faithful witness, and a heart that recognizes the Shepherd’s voice.
The midnight cry will not ask whether our chatbot knew our preferences. It will reveal whether Christ knew us as His own (Matthew 25:1-13). Therefore, let us watch soberly, test carefully, remain in fellowship, and hold fast to the Word of life while we await our blessed hope, the appearing of the glory of our great God and Savior, Jesus Christ (Titus 2:11-13).