Navigating AI for Biblical Engagement

Introduction

Many AI-enabled tools are now marketed to Christians to study Scripture and grow in spiritual maturity. These tools claim that users can ‘create customized study plans tailored to your needs’,¹ ‘engage with an AI-powered friendly theological professor’² or even ‘conduct advanced research on complex biblical topics’.³ How should Christians be viewing AI-enabled tools for biblical engagement? In this piece, we break these tools into three categories. We consider the drawbacks of using such tools for spiritual growth and offer recommendations to congregants and clergy encountering these tools.

Categories of AI Tools

Retrieval-Based Tools

Retrieval-based tools take a user’s statement as input and return search results from a curated database.⁴ For example, a user might be interested in the theme of sacrifice in the Old Testament and could use an AI tool to quickly pinpoint stories like Abraham’s sacrifice of Isaac (Genesis 22) and Elijah’s sacrifice on Mount Carmel (1 Kings 18). For example, the Cross Reference Explorer from VerseOnix allows a user to ‘search any word, theme, or story to explore its connections across Scripture’.⁵ The curated database would include individual verses or passages of the Bible but could also incorporate other resources ranging from patristic texts to recent sermon transcripts. Retrieval is often coupled with large language model (LLM) generated text (known as retrieval-augmented generation, or RAG),⁶ where the original passage is returned alongside some additional explanation.

Pre-generated Textual Tools

Pre-generated textual tools like Open Bible’s AI-Assisted Bible Study⁷ rely on a database of content previously generated using an LLM. Based on a user query, a subset of relevant content will be returned. For example, for a specific passage of scripture the LLM-generated text could include a summary, discussion questions, commentary, or interpretation. The benefit of this approach is that no AI tooling needs to be invoked for a user’s particular search and developers can manually curate entries to make incremental improvements.

Conversational Chatbots

Conversational chatbots are by far the most common tool, where a user dialogs with an LLM. Developers market these LLMs as ‘more Christian’ than their ‘secular’ counterparts like ChatGPT or Claude. Setting aside the problem of quantitatively verifying these claims,⁸ there are several approaches developers can use to customize LLMs based on content like Bible verses or other Christian resources. These customized chatbots are cheap and easy to build and require little to no coding experience. Engaging with these models also raises the most technical and theological problems.

Possible Benefits, Drawbacks, and Recommendations

Enriching Bible Study

If used alongside other reliable sources, RAG-enabled LLMs have the potential to introduce users to a variety of materials they may not have otherwise engaged with. These materials could enable a Christian to grow in their understanding by exploring perspectives that differ from their own. For example, a Christian from a charismatic denomination might be inspired by the English Catholic mystic Margery Kempe. Christians from a Catholic or Protestant background could be introduced to Eastern Orthodox scholars like Kallistos Ware or Alexander Schmemann. Finally, including secular sources in a RAG corpus could lead to a richer understanding of what makes the Christian perspective so distinctive.

Limiting the Hermeneutical Cycle

There are some clear downsides when using an AI model for Biblical engagement or spiritual formation. Leaning too heavily on an AI model for insights has the potential to shortcut meaningful growth. St Augustine wrote about a hermeneutic cycle that generates understanding by oscillating between faith and reason, text and context, initial understanding and new awareness.⁹ In modern hermeneutics, these ideas are reflected in the language of the ‘hermeneutical spiral’¹⁰ and the ‘fusion of horizons’.¹¹ Relying on an LLM for meted-out platitudes can break this cycle of development. Christians are each called to present ourselves to God as ‘a worker who does not need to be ashamed and who correctly handles the word of truth’ (2 Timothy 2:15). Since an AI tool is incapable of handling the word of truth, having no soul beloved by God or made in God’s image, perhaps it is not worth incorporating LLMs into our Biblical engagement.

Cultural Bias

When conversing with an LLM, common problems are heightened when the LLM is used for biblical engagement. Confabulation is a present concern and could lead to misunderstanding, even heresy. Biases across cultural and social dimensions, as well as doctrinal dimensions, are difficult to gauge and may reinforce one’s existing proclivities. Since most LLMs are trained on misappropriated material from Western Christian scholarship, they privilege certain culturally bound interpretations above others.

Anthropomorphizing

Chatbot anthropomorphizing is a growing concern,¹² which in the context of spiritual formation may lead some users to see the LLM as a spiritual companion or guide rather than the powerful yet flawed tool it is. Any one of these problems in isolation should give pause about using an LLM for spiritual formation. Taken together, it is doubtful whether LLMs should have a role in Biblical engagement.

Human Discernment

A common argument in favor of using these tools is the fallibility of human beings: if an AI can exceed the average human in explaining a passage of scripture, it is worth deploying to users. However, as Luke Plant points out, an LLM performing exegesis or offering spiritual direction or evangelizing to a crowd is not competing with the average human, it is competing with the best resources readily accessed on the internet.¹³ Going one step further, an LLM should be no match for the pastoral oversight and community engagement a maturing Christian should already be participating in. The work of biblical scholarship, interpretation, teaching, and learning is the work of the body of Christ, and cannot be responsibly performed by a machine. God is among those where ‘two or three are gathered in my name’ (Matthew 18:20), and the returns of an LLM are no substitute for Christian community and scholarship.

Conclusion

For church leaders encountering these tools, it should be clear that an LLM in isolation should not be trusted to interpret scripture, foster Biblical engagement, or evangelize to unreached peoples. These tasks have been entrusted to us, the church, Christ’s body. They are too important to delegate to a machine. They are too full of living purpose to be left to a dead and rudderless apparatus. Theologian Judith Wolfe wrote, ‘When we construct theologically, we are not building towers; we are building boats. And we trust the sea.’¹⁴ Who would dare venture forth on a flimsy, rudderless skip, full of holes?

Endnotes

1. “BibleAI: Discover the Most Advanced Bible Search Engine”. bibleai.com. Accessed 9 August, 2025.
2. “SharetheBible.ai: Discover Timeless Biblical Insights with Artificial Intelligence and Share the Bible Confidently”.sharethebible.ai. Accessed 9 August, 2025.
3. “ScriptureArk: Advanced Bible Research and Study Tool.” criptureark.com. Accessed 9 August, 2025.
4. Graber, Adam. “Robot ‘Church Fathers’ Might Curate New Cannons”. Christianity Today, 2023.
5.  “VerseOnix: Discover Scripture Through AI-Assisted Study”. verseonix.com. Accessed 9 August, 2025.
6. Lewis, Patrick, et al. “Retrieval-augmented generation for knowledge-intensive nlp tasks.” Advances in neural information processing systems 33 (2020): 9459-9474.
7. “AI-Assisted Bible Study”. openbible.info/labs/ai-bible-study. Accessed 9 August, 2025.
8. Schwarting, Marcus. “To Christians Developing LLM Applications: A Warning, and Some Suggestions”. AI and Faith, 2023.
9. Saint Augustine. “The Confessions of S. Augustine”. Vol. 1. (JH Parker:1838).
10. Osborne, Grant R. “The hermeneutical spiral: A comprehensive introduction to biblical interpretation.” InterVarsity Press, 2010.
11. Gadamer, Hans-Georg. “Truth and method”. A&C Black, 2013.
12. Reinecke, Madeline, et al. “The Double-Edged Sword of Anthropomorphism in LLMs”. MDPI Proceedings, 2025.
13. Plant, Luke. “Should we use AI and LLMs for Christian Apologetics?”. 2024.14. Judith Wolf, The Theological Imagination: Perception and Interpretation in Life, Art, and Faith. (Cambridge, UK: Cambridge University Press, 2024) 26.
14. Judith Wolf, The Theological Imagination: Perception and Interpretation in Life, Art, and Faith. (Cambridge, UK: Cambridge University Press, 2024) 26.