Should Christians use AI?
Over the past two years, I have been asked this question in many forms. Sometimes it is very practical: Is it safe to use ChatGPT for Bible study? Should pastors use AI to prepare sermons? Sometimes the concern goes deeper: as AI becomes more powerful, could it eventually replace pastors, or even the church?
Attitudes vary widely. Some Christians have embraced AI enthusiastically; others remain deeply suspicious. Many, I suspect, live somewhere in between. They already rely on AI at work or casually ask it to plan a holiday or organize information, but when the conversation moves towards Scripture, faith, or spiritual formation, they hesitate. I understand that hesitation. We have been asking many of the same questions while developing Sparks, a digital initiative of the British & Foreign Bible Society.
Sparks (www.sparks.bible) did not begin as an AI project. What prompted it was an older and more fundamental missional question: once people have access to the Bible, what helps them enter more deeply into Scripture and allow God’s Word to shape their lives? Several years ago, I described this in ChinaSource as the “last mile” of Bible engagement.
After generations of Bible translation, publishing and distribution, followed by the rapid growth of digital technology, we have access to biblical resources on a scale previous generations could hardly have imagined. Yet having a Bible is not the same as reading it; reading is not necessarily understanding; and even understanding does not mean that Scripture has entered a person’s life deeply enough to shape their judgment, choices, and character. Sparks began as one attempt to address that last mile.
For ChinaSource readers, a brief word of context may be helpful. Sparks primarily serves Chinese-speaking Christians across different countries and regions. It is developed and operated outside mainland China, where most of its users are currently based, while remaining accessible from within mainland China.
Our original idea was fairly straightforward: build a digital Bible resource and learning platform where trustworthy resources, structured learning and practical engagement with everyday life could be brought together in one place. I often imagined it as a library. The problem is that libraries take time to build. Even for a Bible Society, good content has to be found, commissioned, produced, licensed, edited and organized, while institutional processes take their course. The time and coordination required turned out to be far greater than we had imagined. Eventually, we had built a useful little library. But rather than the national library we had imagined, it felt more like a good neighborhood library.
Then generative AI arrived. Initially, we thought perhaps this is the tool we have been waiting for.
What Changed with AI?
For Bible engagement, the possibilities were obvious. AI could make biblical knowledge far more accessible, conversational and responsive than a conventional digital resource library ever could. But questions quickly followed. Where do these answers come from? Whose theology is being presented? Can the answers be trusted? Some of these concerns reflect genuine technical limitations—hallucination, accuracy and provenance—and they remain important even as the technology advances rapidly. Yet the more we worked with AI, the more a different kind of question came into view.
Even as AI becomes more accurate and better grounded in trusted sources, it cannot decide for us which interpretations carry theological authority. Nor can technical improvement answer a more fundamental question: what place should AI occupy in the formation of a Christian? As the technology became more capable, these deeper questions became more pressing. And so, within Sparks, we began asking a more provocative question: what if the church does not need another Christian AI?
Not because AI is not good enough. Quite the opposite. Precisely because it is becoming so capable, we need to be clearer about what we are actually asking it to help us do.
Sometimes What I Need Is Not a Correct Answer
Suppose I have been treated unfairly by a colleague. I open an AI tool and ask, “What does the Bible say I should do?” Within seconds, it can probably identify passages about anger, justice and forgiveness, explain them coherently, and offer several practical suggestions. The answer may be entirely correct, and it may genuinely help me. But I may still be furious.
More importantly, everything the AI knows about the situation has come from me. I may describe my colleague’s behavior in great detail while quietly minimizing my own responsibility or be very receptive to what Scripture says about justice while somehow passing quickly over what it says about forgiveness. The AI may then take my version of the story and construct an impressively coherent—even apparently biblical—response around it. At that point, the real question is no longer whether AI has given me the correct information. It is whether I am willing to be corrected.
Perhaps what I need is not another answer, but to remain with a passage of Scripture long enough for it to begin questioning me. I may need a friend who knows me well enough to say, “Arleen, I don’t think that is quite the whole story.” I may need the accumulated wisdom of the church, the experience of others, and time to pray. In the end, what I may really need is the courage to go and have an honest conversation with my colleague.
This is where our understanding of Sparks began to change. Initially, we were mainly asking how technology could help people understand the Bible better. Increasingly, we found ourselves asking what kind of environment helps people to be formed through their encounter with Scripture. The two questions are related, but they are not the same.
From Information to Formation
Digital technology is very good at information. We search, click, and consume. Algorithms learn what interests us and send us more of what we are likely to want. Generative AI takes this much further: it can reorganize information around me—my language, my level of knowledge, my question, my situation. Christian formation does not work entirely according to that logic.
Scripture does not always orbit around my questions. Sometimes it unsettles them. I come asking, “How should I deal with this difficult person?” and discover that the text is beginning to address my anger. I come looking for an answer and find myself drawn into a story that I am called to live. The direction changes. Instead of only asking, “How can I use Scripture to interpret my life?” I begin learning to allow Scripture to interpret me.
This has increasingly become central to what we are designing Sparks to become. We no longer want Scripture to be simply one resource among many, accompanied by some Christian content and an AI chatbot. We want to build an environment that helps people remain in sustained encounter with Scripture. AI has a place in that environment, as do theological resources, lived experience, relationships and the wisdom of others. But all of them should deepen our encounter with Scripture rather than replace it.
Internally, we have begun calling this a formation ecology. The idea is simple: formation is never accomplished by a single tool. Scripture, community, relationships, tradition, experience, and technology each have a place. AI can be part of that ecology, but it cannot be the whole of it.
When the Mission Changes, Building Changes Too
This realization has made Sparks more difficult to build, not less. If Sparks were simply a digital Bible library, the task would be relatively clear. If it were simply a Christian AI, the technical challenge would also be easier to define. But a Scripture-centered environment aimed at Christian formation requires something much broader.
It needs biblical and theological understanding alongside product design and technology and knowledge of formation and community alongside research and operations. It needs the trust and long-term responsibility that established Christian institutions can provide, but also the experimentation and room for adjustment that innovation requires. Those capabilities rarely sit naturally in one team or organization.
Projects like this can also attract considerable support at the level of ideas while struggling to secure the practical commitment needed to build them. Money matters, but so do professional expertise, time, partnerships, and a willingness to accept the uncertainty that accompanies innovation. This has become part of what Sparks is teaching us: the challenge is not simply to invent a better technology, but to bring together the people, capacities and resources needed to serve a much larger mission.
I should also be candid about where we are today. If you visit Sparks now, you will not yet see the full picture I have described. It still looks much closer to our original conception: a digital Bible and resource platform with an AI chatbot. The product has not yet fully caught up with our deepening understanding of the mission.
That gap can be uncomfortable, but it has also been instructive. We began building around one understanding of the problem. Building itself, and watching how people used what we built, helped us see that the problem went deeper. We now have a clearer sense of where we want to go: not simply towards a platform where people find biblical content or ask AI questions, but towards an environment in which Scripture, community, lived experience, Christian tradition, and technology each find their proper place in serving Christian formation. And that has changed not only what we want to build, but how we think it should be built.
Building “With,” Not Just “For”
If formation happens through relationship and participation, can we design the entire environment behind closed doors, launch a finished product, find a group of “users”, and then announce that we have created a community? Increasingly, we think not. Nor can we assume that we already know all the answers and then invite people to participate in something we have essentially finished designing.
Sparks itself needs to take shape through the participation of people with different gifts, experiences and forms of knowledge. So we are moving from building for a community towards building with one. We call the framework supporting this approach VOICE: “Every voice. One Word.” There is a deliberate tension in that phrase.
Every voice matters. People are not passive objects waiting to be “formed”; they bring stories, experience, questions, doubts and different ways of hearing Scripture. But that does not make every voice the final authority. Scripture remains our anchor. Co-creation therefore does not mean voting on what is true. It means creating spaces where lived experience comes into genuine encounter with Scripture and with other people—and allowing those encounters to shape what we build.
We do not yet have all the answers about how to do this well. Perhaps that is precisely the point. A formation community is not something we can fully engineer in advance.
Perhaps We Need a Better Question
So, should Christians use AI? I think we should explore it—with wisdom, critical judgment, and imagination. The possibilities AI opens for Bible engagement are too significant to ignore, particularly for people who lack access to trusted resources or pastoral support.
But the question I increasingly find myself asking is not whether Christians should use AI. It is what kind of people and communities we are called to become, and where technology should stand within that process. The church may not need another “Christian AI,” but it does need to reimagine how, in an age of AI, we build environments around Scripture that genuinely contribute to the formation of people.
Doing that will require more than better AI or a better app. It will require new partnerships, new combinations of expertise, new ways of resourcing innovation, and communities willing to learn and build together. Perhaps when we evaluate Christian technology, we should ask not only what we built, but what kind of community we are becoming as we build it.
The “last mile” of Bible engagement is not something technology can travel on our behalf. We have to travel it together.
Every voice. One Word.
