AI, the Bible, and our Future
5.1 Using AI as a Christian Responsibly
image generated with FLUX-pro on Poe
Last Updated: November 30, 2025
Stewards
God has made us, as humans, stewards of this world. You have a responsibility for the things the Lord lays directly into your hands. Consider the gospel message that has saved you, your workplace, the money you receive, and the things you own. None of these things truly belong to you, for “the earth is the LORD's, and the fulness thereof; the world, and they that dwell therein” (Psalm 24:1). We will surely be held accountable for our stewardship of them (Matthew 25:14-30). We can steward them for His glory or for the glory of man, but the latter will be brought to nothing before Him (1 Corinthians 1:29).
AI is no different: It is a technology that we as Christians must steward well. Whatever we have, we ought to steward it to the glory of God (1 Corinthians 10:31). The first question we must therefore ask ourselves is: “Can AI be used for the glory of God?”
1 Corinthians 10:3131 Therefore, whether you eat, or drink, or whatever you do, do it all to the glory of God.
Scripture does not forbid the use of technology. Since the beginning, mankind has made and used tools. Think about the papyri the apostles used to write letters to the churches or the printing press in Luther's time that made the Bible accessible to all! Even beyond these obvious examples, consider King Uzziah, who built war machines (2 Chronicles 26:15). God’s word says that he did what was right in the eyes of the Lord (2 Chronicles 26:4), and it was only after building those machines, when pride entered his heart, that he sinned against God (2 Chronicles 26:16). So we see that the building of the machines was not the issue; the technology itself was not inherently evil. It is the heart of man that is deceitful and wicked (Jeremiah 17:9).
The next question, therefore, is: “How can we make use of AI for the glory of God?” The answer is simple if we understand God’s will. We glorify God by using AI in a way that draws us and others closer to Jesus Christ. We glorify Him by depending on Him, being conformed to His image, and giving thanks in all things (1 Thessalonians 4:3; 5:18).
AI for Bible Study
One of the most obvious ways this can be achieved is by using AI in your personal Bible study.
This might be done by letting the AI…
- explain a passage in the Bible or generate commentary on it.
- provide cross-references.
- find paraphrased Bible verses that are difficult to locate with a standard search engine.
- provide details on the Hebrew, Aramaic, and Greek original manuscripts.
- provide detailed historical and cultural context for Bible passages.
The list above is certainly not exhaustive but should give a good overview of how AI could be used for your Bible study. But before you use it at all, you should understand how it works, at least on a basic level. Understanding how it works is necessary to know its capabilities and limitations, which is the key to using it responsibly.
How AI Works
What is often simply referred to as “AI” is usually a very distinct AI model called a “Large Language Model” (LLM). This is the technology behind AI chatbots where you write a message and receive a response. ChatGPT, Claude, Gemini, and Grok are some examples of such LLMs (though, at this point we usually have what is referred to as “Multi-Modal Language Models” (MMLM) which are able to process not just text, but several other modalities such as audio, image, and even video).
To create an LLM, a huge artificial neural network, vaguely inspired by the human brain, is trained on a vast amount of data from the internet. During this training, the AI model learns the probabilities of one word (or more precisely, a "token," which is often a part of a word) following another in a given context.
Imagine you have sent an article to ChatGPT. As it generates a response, it builds the text one token at a time. Let's say it has generated the phrase: “Interesting article about this.” Now, let's look at what happens before the next word is chosen:
This screenshot is taken directly from the OpenAI playground, where you can test different LLMs and as you can see, the most likely token to follow is “is” with a probability of 9.14%. The next most likely is “project” with a probability of 6.11%, followed by a line break (3.83%) and “dataset” (2.13%).
A parameter called "temperature" adds randomness to this selection process. When the temperature is set to 0, the model always picks the token with the highest probability—in this case, “is.” But if you set the temperature higher, the AI introduces a bit of randomness, possibly choosing from the top three, five, ten, or even one hundred most likely tokens. It would still choose “is” most of the time, but it might sometimes choose “project,” the line break, or “dataset.”
After a token is picked, it is added to the output text, and the new, longer sequence is sent back to the model to again calculate the probabilities for the next token. This process repeats until the AI model predicts a special “end-of-text” (EOT) token, which signals that the response is complete.
Initially, the AI model can only continue an input text, much like an highly advanced auto-complete. This "base model" is not yet a chatbot. In a second training stage, called "instruction-tuning," the model is trained to follow instructions and respond helpfully. Additional training stages are required to refine the model into a capable chatbot. Through this complex process, an emergent capability arises that can be accurately described as "Artificial Intelligence."
It is crucial, however, to make a clear distinction between “intelligence” and concepts like “consciousness,” “feelings,” or “intentions.” None of the latter could ever truly be found in an LLM or any other AI model, even if it claims to possess such things. I believe the Bible makes it clear that these attributes of life are given specifically by God, according to His sovereign will, not by the will of man. Nevertheless, LLMs do possess an ever-increasing intelligence and are becoming more sophisticated and general in their capabilities.
You might be wondering what all of this has to do with using LLMs for Bible study. Despite their growing intelligence, a key limitation of current LLMs is that they do not possess an accurate sense of truth. They don't "know" what they know or don't know. They simply respond with the most statistically likely text completion that fits the given context. In this process, LLMs often “hallucinate.” This is a technical term for the phenomenon where an LLM generates information that looks convincingly true on the surface but is actually false or nonsensical. These hallucinations are the critical reason why LLMs may not always be factually correct and are sometimes completely mistaken in their output.
Practical Application
Practically, this means that LLMs might sometimes…
- cite Bible verses incorrectly (a very real issue for nearly every Bible translation, except the KJV, which has around 98% accuracy in citing Bible verses correctly).
- provide false or nonsensical cross-references.
- make up information about the original manuscripts and the meaning of Hebrew, Aramaic, and Greek words.
- teach false doctrine (this too is a very real issue, especially concerning the baptism in the Holy Spirit).
Therefore, whenever you use an LLM in your Bible study, you must: read Scripture for yourself, critically evaluate the truthfulness of the AI's statements, and always take its output with a grain of salt, especially when it concerns people or historical claims. If you intend to use the AI's output for making judgments or sharing information publicly, you must first verify its claims with reputable sources.
Dangers
Within all of this, there are two grave dangers: 1. Dependency, and 2. Seeking knowledge rather than revelation.
Firstly, an LLM must never be a replacement for your dependence on the Holy Spirit, for He is the one who guides us into all truth (John 16:13). Very quickly, an LLM could become your personal prophet or oracle, telling you what you ought to do in every situation, and slowly but surely distance you from Christ and damage your personal relationship with Him. Be on guard, question yourself, speak with your brothers and sisters in Christ, and do not separate yourself from fellowship or the teaching of the Word (Acts 2:42; Hebrews 10:25).
Secondly, it seems to me that there are three stages of knowing: 1. Memorization, 2. Understanding, and 3. Revelation. Simply knowing something to be factually correct is not enough (Ephesians 1:17-18). LLMs can greatly aid you in memorizing biblical truths (for example, the order of creation in Genesis) and even understanding them more deeply, but they cannot reveal those truths to you! This is something only the Holy Spirit can do in you. You must go to prayer and seek the face of the Lord to behold Him in His beauty (Psalm 27:4) and know Him personally! Never, no never, be dependent on anything but the Word of God and the Holy Spirit (John 15:1-6), especially in getting to know Him personally and growing in your relationship with Him, through which He sanctifies you (Leviticus 20:8; 1 Corinthians 1:30).
Thirdly, it is of paramount importance to understand that AI can never give you practical advice in wisdom. AI sure can give you earthly wisdom, but not the wisdom you really need. True wisdom is from above and comes from God (James 3:13-17). This wisdom is imparted to us by the Holy Spirit—humans, not AI. If you have a real spiritual need, don’t go to an AI model, go to a brother or sister in Christ.
Beyond that, you must ensure that you do not spread misinformation. If you simply repeat something an LLM said without having judged and checked it yourself, you will surely be held accountable for it before the judgment seat of Christ (2 Corinthians 5:10).
Final Recommendations
Be prayerful about your AI usage, not just for Bible study but in every area of your life. The Lord alone knows in which ways you can safely and responsibly use AI for His glory. He will help you and show you the way if you seek Him for help (1 Chronicles 28:9).
The Bible Study Tool
The Premises
For the Bible Study Tool, we need to establish a foundation on which the AI chatbot should generate its responses. Within Scripture we can find a set of premises:
Premise 1
The entire Bible is the inspired, inerrant Word of God (2 Timothy 3:16).
2 Timothy 3:1616 All Scripture is God-breathed and profitable for teaching, for reproof, for correction, for training in righteousness,
Premise 2
Proper biblical interpretation requires diligent study of the original Hebrew and Greek texts, understanding historical, cultural, and scriptural context, cross-referencing within Scripture, and comparing Scripture with Scripture (2 Timothy 2:15).
2 Timothy 2:1515 Be diligent to present yourself approved to God as a workman who does not need to be ashamed, accurately handling the word of truth.
Premise 3
God's Word is the highest authority and truth on all matters it addresses - science, history, morality, etc. Human knowledge and reasoning must be subservient to Scripture (John 17:17).
John 17:1717 Sanctify them by the truth; Your word is truth.
Premise 4
A literal, historical-grammatical hermeneutic should be employed for the interpretation of Scripture. Consider for example Exodus 20:11: How should we interpret that verse?
Exodus 20:1111 For in six days Yahweh made the heavens and the earth, the sea and all that is in them, and rested on the seventh day; therefore Yahweh blessed the sabbath day and made it holy.
The hermeneutical approach emphasizes understanding Scripture in its plain, ordinary sense within its historical and grammatical context. It interprets the Bible as the original authors intended and as the original audience would have understood it, considering the historical setting, cultural background, and the normal rules of grammar and language usage of the time. This method assumes that if the plain sense of the text makes sense, we should seek no other sense (in the case that we would be rejecting the true meaning and adding our own ideas into the text—essentially performing eisegesis), while still recognizing and properly interpreting figures of speech, symbolism, and literary devices as they would have been understood in their original context. This approach leads to a straightforward reading of Scripture, including passages like Exodus 20:11, which presents the creation account as six literal days, thereby anchoring our interpretation in the text's historical claims.
Premise 5
The Bible is a unified, internally consistent revelation from God. Apparent contradictions stem from our limited understanding, not actual errors (2 Peter 3:16; Psalm 119:160).
2 Peter 3:1616 as also in all his letters, speaking in them of these things, in which are some things hard to understand, which the untaught and unstable distort, as they do also the rest of the Scriptures, to their own destruction.
Psalm 119:160160 The sum of Your word is truth, and every one of Your righteous judgments is everlasting.
The System Prompt
Every chatbot has a so-called System Prompt. This system prompt specifies the behavior it should adopt and the format of its responses. In the chatbot that I made on the platform Poe (you can find it as Bible_Study_Tool), I specified those exact premises that I outlined above. Following the premises, I outline two more Biblical principles to reinforce the premises. In a section for operational guidelines, I specify for example the need to respond in the user language. At the very end I specify one fundamental doctrine it still gets wrong quite often because there is so much trash about it on the internet. It is the one I already mentioned before:
There are two kinds of Spirit baptism:
- Baptism by the Spirit into the body of Christ, which happens at the new birth (1 Corinthians 12:13)
- Baptism by Christ into the Spirit, which is a subsequent experience (Mark 1:8; Acts 1:4-5.8; Acts 2:1-4.33)
If you believe to have found any serious problem with the Bible Study Tool, feel free to reach out to me by writing a comment.
With all of the above in mind, you should be well equipped to use the AI chatbot responsibly. If you want to know how to prompt the LLM most effectively, you can read my blog post 5.2 Prompt Engineering!
If you found all of this to be helpful, consider subscribing to my blog for getting updates or sharing this post!
All Bible quotes are from the Mondern English Version (MEV) accessed via BibleGateway https://www.biblegateway.com/passage
Thanks
Great tool.
Formatting with markdown would take it to the next level, so users aren’t staring at a wall of text.
Thanks.
That’s actually a great idea! I will make some tests if it influences biblical accuracy. If all works well, I will implement it in the system prompt!
Very well presented, brother. There is none of the Life of God (John 1:4) in AI, therefore we should not consult it as though it is a Living contributor like we would a brother or sister in the Lord.
Great article Jonathan! Thank you!