AI is a powerful tool that is becoming part of everyday life, including religious life. In a May 2026 report, Barna found that 30% of surveyed U.S. adults agreed that spiritual advice from AI is as trustworthy as advice from a pastor.
A Christian who asks AI what the Bible says may be trying to understand a passage. They may be looking for guidance on a difficult decision and wondering what God asks of them. A person from another faith tradition may bring similar sorts of questions about how to live in relation to God and neighbor. In each case, people are turning to AI as an authority to answer their questions faithfully, accurately, and without hallucination.
These uses call for discernment about where AI helps and where it should not be used. Faith communities need the language to question what a product promises, set boundaries, and refuse uses that undermine their responsibilities. By exercising this discernment and calling for specific change, people of faith can have a role in shaping this technology.
All people should expect AI to explain where its answers come from, be honest about its limits, and leave them free to question its advice. They should also know who can help when something goes wrong. Faith communities uniquely have a say in how those expectations are tested and if AI companies want greater trust, they need to show why it is warranted.
Consider a situation of someone asking an AI about forgiveness and whether they should resume a relationship in which they have been harmed. An answer that sounds compassionate may press someone toward a decision an AI system is unqualified to make. Just accurately describing a religious teaching does not establish an ability to offer responsible personal counsel.
Represent beliefs faithfully
Faith and spiritual life questions can involve a sacred text, an inherited practice, or the judgment of a trusted teacher. An assistant needs to make clear where its answer comes from. If it quotes a text, the words should match the source. If it offers an interpretation, readers should be able to distinguish that explanation from the teaching it claims to represent.
Disagreement matters too, including within a tradition. An assistant that blends different views into one confident answer may leave someone unable to tell whose guidance they are receiving. Developers should explain which sources and interpretations shape the system’s responses and how they handle disagreement. A label such as “faith-based” tells a user very little on its own.
Help people exercise judgment
An AI assistant could help someone study a piece of text, understand a practice, or prepare questions for a religious teacher. The question is whether it helps them understand the answer and its limits, consider alternatives, and make their own decision. They should remain free to question the advice or refuse it.
In the Christian contexts examined by Fide’s early work, spiritual formation is informed by growth in faith, character, and love for others over time. Other traditions should define the purposes and practices that matter to them. More frequent use of an app or a high score on religious questions doesn’t mean that a person is truly being spiritually formed. Some of the most valuable assistance may help someone put the device down and speak with another person, return to a text, or take time to think.
People must also be able to question religious leaders or seek help outside their community. An AI assistant should not dismiss their concerns simply because someone in authority has told them what to do.
Give communities evidence they can use
Faith institutions should expect developers to explain what their systems have been tested for, where they fail, and what remains unknown. People need to know what happens to sensitive conversations, how to report concerns, and who can correct or withdraw a harmful feature. Institutions choosing to use AI have responsibilities here, but should not inherit the entire burden of discovering a product’s failures.
Religious practitioners, scholars, and people seeking guidance can help identify what needs examining. Researchers can work with them to design tests that developers can learn from. That requires room to disagree about what a good response looks like. Participation should not depend on sharing private conversations or accepting one institution as the voice of an entire tradition.
Our Work at Fide AI
Fide AI brings measurement science and technical rigor to the concerns and experiences of high-trust domains like faith communities. That means turning questions about religious guidance, trust, and authority into tests that reveal what AI systems do, where they fail, and which changes improve their behavior.
The first studies examine how AI responds to Christian questions and situations requiring human care, quotes Scripture accurately, and uses trusted sources when users ask it to skip them. The findings show how instructions and the surrounding software affect responses, while identifying failures that remain. They do not establish that AI can provide pastoral care or help someone grow spiritually.
The goal is to give developers evidence they can use to improve the models and products being built today and to guide future research. Faith communities can help shape that work by identifying the questions that matter in their settings and what responsible assistance should look like.
Questions to ask before using AI in your community
Start with a tool your community already uses or is considering. Describe one realistic situation where its answer could matter, leaving out private details. Ask the provider how it tested that situation, what failed, and who responds to concerns.
The company providing the tool is a practical first contact. The labs developing the models behind it and independent researchers may also be useful partners. Researchers can help turn a concern into a test; you can contribute knowledge of the situation without needing to design the evaluation yourself.
Use these questions to guide the conversation:
You should be able to get answers in plain language, backed by evidence. If an important answer is missing, that is a reason to limit the tool’s role until the concern is addressed.
The questions above give communities a way to ask for that evidence and give researchers concrete problems to investigate.
If you want to learn more, explore the research or if you’d like to contribute reach out or share a faith-domain area of study with Fide AI.
Sources
Barna reports surveyed U.S. adults' views on the trustworthiness of AI spiritual advice. Three Christian-context Fide AI studies illustrate the research approach; their findings do not establish results across faith traditions.