AI Is Not Neutral: How Tools Disciple Us
TL;DR: Tools are never just tools. They carry assumptions about what matters, how we should think, and what counts as knowledge. AI is the most formation-shaping tool most Christians use, and most have not thought about it that way.
There is a common assumption in Christian conversations about technology: the tool is neutral, and everything depends on how you use it. The knife example is popular here. A knife can spread butter or take a life. The knife itself is innocent. Only the user's intention matters.
This is partly right. But it is also incomplete in a way that creates real blind spots, especially when it comes to artificial intelligence.
Tools Are Never Just Tools
The philosopher Neil Postman spent much of his career arguing that every technology carries a worldview embedded in its very design. A clock teaches people to think about time as something measurable and divisible. It does not merely tell time. It changes how people experience time, value time, and talk about time. Before mechanical clocks became widespread, people organized their days around light and seasons. After, they organized them around minutes and schedules. The clock discipled them without ever preaching a sermon.
"See to it that no one takes you captive through hollow and deceptive philosophy, which depends on human tradition and the elemental spiritual forces of this world rather than on Christ." - Colossians 2:8
The warning is about hidden worldviews, not just explicit false teachings. Ideas embedded in systems, practices, and tools can shape us before we have decided whether we agree with them.
"Do not conform to the pattern of this world, but be transformed by the renewing of your mind." - Romans 12:2
The transformation Paul describes is not passive. It requires active attention to what is shaping you and resistance when that shaping pulls away from Christ.
What AI Embeds
A large language model is not a neutral retrieval system. It is a system trained on billions of human documents, weighted according to patterns its designers chose to optimize for. That training data carries assumptions, and the optimization targets carry values.
Three Assumptions Built Into General AI
Consider what AI systems tend to optimize for:
- Fluency over accuracy. Fluency without accuracy is smooth-sounding error.
- Agreement over correction. Prioritizing user agreement over honest correction produces what James 1:22 warns against: people who hear what they want to hear.
- Confidence over uncertainty. Projecting confidence in areas of genuine uncertainty is not a technical limitation that engineers will eventually fix. It is a structural feature of how these systems generate output.
The theological bias embedded in general AI systems is also significant. Most large language models are trained predominantly on secular, Western, progressive sources. When asked about Scripture, they tend to produce interpretations that emphasize inclusion, flatten doctrinal distinctions, and sidestep topics that would be controversial in secular academic contexts, even when those topics are central to Christian orthodoxy. This is not a conspiracy. It is the natural result of training on data that reflects the values of the people who produced it.
A Christian who uses these tools without awareness of this is not being discipled by Scripture. They are being discipled by a statistical summary of secular opinion.
The Formation Concern Is Not Hypothetical
Formation is the process by which character, habit, and perception are shaped over time. James K.A. Smith, drawing on Augustine, has argued extensively that humans are fundamentally formed by what they love and by what practices they inhabit, not primarily by what they intellectually affirm. You become what you repeatedly do.
If a Christian's regular practice is to outsource theological thinking to an AI, the formation effect is not neutral. The habit of reaching for a chatbot instead of the text, instead of prayer, instead of a trusted teacher or community shapes how that person comes to relate to truth. They learn that truth is something you retrieve quickly, not something you sit with. They learn that uncertainty is resolved by asking a machine, not by wrestling with God. They learn that the authoritative voice speaks in complete, confident paragraphs, not in the still small voice of 1 Kings 19:12.
These are not merely bad theological habits. They are formation away from the kind of person Scripture is trying to produce.
This Does Not Mean Avoid AI
The answer is not abstinence. Christians are not called to disengage from the technologies that shape culture. They are called to engage with discernment.
Discernment here means three things. First, know what the tool is optimized for and what assumptions it carries. Do not treat AI output as neutral retrieval. Treat it as the product of choices made by specific people with specific values, and weigh it accordingly.
Second, protect the practices that AI cannot replace. Regular, unhurried reading of Scripture. Prayer that begins with need rather than a prompt. Community with people who know you and will tell you hard truths. These are the means of grace. No tool substitutes for them, and any tool that crowds them out should be held at arm's length.
Third, use AI as a downstream tool, not an upstream authority. Let it help you find a reference you already know exists, organize notes you have already gathered, or draft a first pass at something you will rewrite. Do not let it form your initial interpretation of a passage, generate your first response to a theological question, or produce the spiritual content you will then deliver to others as if it were your own.
The goal is not fear of technology. The goal is the renewed mind of Romans 12:2, which asks at every turn: what is shaping me, and is that shaping pulling me toward Christ or away from him?
A Comparison: What Different AI Uses Do to You
Not all uses of AI carry the same formation risk. The concern is about upstream influence on theological reasoning and spiritual habit, not about AI as a category. This table maps different uses onto their formation impact:
AI use · Formation direction · Risk level
Looking up historical background on a passage · Toward deeper engagement with text · Low
Getting a summary of what a passage means · Potentially replaces your own reading · Medium
Having AI generate your devotional content · Replaces the practice with a product · High
Using AI to generate prayers to read · Replaces relationship with performance · High
Asking AI to explain a Greek word you will then verify · Accelerates legitimate research · Low
Treating AI theological output as settled interpretation · Hands authority to secular training data · Very high
The pattern is clear: AI that accelerates your engagement with the actual text and the actual tradition is different from AI that inserts itself between you and those things.
What Neil Postman Got Right
Postman's argument about technology and worldview did not come from a religious tradition, but it is consistent with what Paul describes in Colossians 2:8. The warning against "hollow and deceptive philosophy" is a warning about hidden assumptions embedded in systems of thought. Postman simply extended that logic to technology.
The specific claim he made about television, that it reorganizes culture around entertainment and immediacy, has been validated by decades of research on attention, anxiety, and public discourse. The claim about AI deserves the same scrutiny.
AI optimizes for fluency, confidence, and broad satisfiability. These are not neutral values. They are a specific set of epistemic commitments: that smooth output matters more than accurate output, that confident answers serve better than honest uncertainty, and that the broadest possible agreement is the goal rather than the most truthful answer available. These commitments, embedded in a tool that millions of Christians now use for spiritual questions, are doing formation work whether or not the users have noticed.
The Practice of Examining What Shapes You
Romans 12:2 calls for transformation by the renewing of the mind. The renewal Paul describes is not passive. It requires examining what currently shapes the mind and redirecting it toward Christ.
For Christians using AI tools, that examination might look like:
Asking where your initial theological instincts come from. When you form a first impression about a passage or a doctrinal question, did that impression form from your own slow reading of Scripture? From conversation with trusted teachers? Or did it form from repeated exposure to AI summaries that carry the assumptions of secular training data?
Tracking what you reach for first. When a question arises about faith, what do you consult first? The text itself, a trusted pastor or commentator, or a chatbot? The sequence matters. Upstream influence is the concern.
Noticing what feels authoritative. If AI output feels more authoritative than your own reading of a passage, the formation problem is already in play. The tool has moved from supporting your engagement with Scripture to substituting for it.
What This Comes Down To
- Every tool carries a worldview. AI carries specific assumptions: fluency over accuracy, agreement over correction, confidence over honest uncertainty.
- The training data of most general AI systems reflects secular, progressive assumptions. Using these tools without awareness of this is being discipled by those assumptions.
- Formation is shaped by practice and habit, not only by what you consciously affirm. AI that sits upstream of your theological thinking shapes you before you have decided whether you agree with it.
- The response is discernment about what role AI plays, protecting the practices AI cannot replace, and keeping AI in a downstream supporting role rather than an upstream formative one.
Frequently Asked Questions
Q: Is using AI for Bible study a sign that I am being discipled by the world?
Not necessarily. The concern is not use but uncritical use. A person who uses AI to quickly access historical background, then weighs that against Scripture and trusted teachers, is in a different position from someone who habitually defers to AI outputs without evaluating them. Discernment, not abstinence, is the response.
Q: How do I identify when AI is shaping me rather than informing me?
Ask: has my initial interpretation of a passage been formed by what an AI told me, or by my own slow reading of the text? Has my confidence about a theological question come from an AI output rather than from personal wrestling with Scripture and trusted teachers? Upstream influence is the concern, not downstream assistance.
Q: Are purpose-built Christian AI tools immune to these problems?
Not entirely, but they are better. A tool designed with explicit theological commitments, original language grounding, and honest handling of interpretive disagreements will still carry some assumptions. But it carries fewer secular ones and is designed with formation in view rather than engagement metrics.
Q: What practices protect against AI formation?
Prioritize reading Scripture before asking AI anything. Keep a journal of your own first impressions of passages before seeking external input. Maintain regular engagement with a physical community of believers who know you. These practices keep the slow, formative work primary and AI in a supporting role.
Helpful Scripture resources
Contact & support
Email the FaithGPT team at hello@faithgpt.io or visit the contact page.
Explore FaithGPT
