In the Bay Area, where church attendance can be as much about schedules and traffic as it is about faith, Justin Lester has been trying to solve a problem that many pastors quietly recognize: people don’t always need spiritual guidance at convenient times. They need it after work, late at night, on weekends, in the middle of a crisis, or simply when a question won’t let them sleep. For years, Lester—pastor of Friendship Baptist Church—has answered those needs as best he could, building relationships through sermons, counseling sessions, prayer circles, and the kind of follow-up that turns a congregation into a community.
But as his church’s reach expanded, so did the gap between what people wanted and what one human being could realistically provide. Lester began to see a pattern: requests for prayer and conversation were arriving faster than he could respond, and many of them came from people who weren’t necessarily able to show up in person. The result wasn’t a lack of care—it was a limitation of time.
So Lester did something that would have sounded futuristic even a few years ago: he created an AI “twin” designed to talk with people about God at any time.
The concept is simple to describe and complicated to build. Lester’s AI duplicate is meant to function like a conversational extension of his pastoral presence—answering questions, offering scripture-based reflections, encouraging users to pray, and guiding them toward next steps. It is not presented as a replacement for the church or for human relationships. Instead, it is positioned as a bridge: a way to keep spiritual conversation available when the pastor isn’t.
And according to reports, the twin is now being used more and more each day.
What makes this story stand out isn’t only the technology. It’s the way Lester frames the purpose. In interviews and discussions around the project, he has emphasized that the goal is not novelty or automation for its own sake. It’s continuity—helping people feel that their questions are heard, even if the answer comes in a different form than they expected.
That distinction matters, because AI in religious settings can easily drift into something else: a substitute that feels too polished, too certain, or too detached from the messy reality of human suffering. A machine can sound confident without understanding the stakes. It can produce comforting language without knowing whether the person on the other end is safe, supported, or in danger. It can also inadvertently steer people away from professional help when they need it most.
Lester’s approach, at least as described publicly, tries to treat the AI twin as a tool with boundaries rather than a free-floating oracle. The twin is built to speak in a pastoral tone and to draw from religious material, but it is also intended to encourage users to connect with real people—especially when circumstances require more than conversation. In other words, the AI is meant to extend access, not replace accountability.
Still, the question many readers will immediately ask is the one that sits beneath every “digital twin” story: how does an AI duplicate someone without becoming a caricature?
A human pastor is not just a voice. He carries a history—his theology shaped by study and experience, his instincts shaped by countless conversations, his empathy shaped by grief and joy. An AI system can mimic patterns in language, but it doesn’t naturally carry the same moral weight behind those words. That’s why the creation of a twin raises a deeper issue than accuracy. It raises authenticity.
Lester’s project appears to be built around the idea that authenticity can be partially engineered: the AI can be trained to reflect a specific pastoral style, to use scripture responsibly, and to maintain a consistent mission. But authenticity is also relational. It depends on trust, and trust depends on transparency. If users believe they are speaking to “Justin” in a literal sense, the ethical ground shifts quickly. If they understand they are speaking to an AI designed to emulate his guidance, the relationship becomes more honest—more like a mediated conversation than a direct encounter.
This is where the twin’s success may hinge less on how well it imitates Lester and more on how clearly it communicates what it is.
Another challenge is the difference between answering questions and providing care. Spiritual care is not only information; it’s timing, discernment, and the ability to notice what a person is not saying. A pastor hears the tremor in a voice, recognizes when someone is spiraling, and knows when to slow down. An AI can detect keywords and sentiment, but it cannot fully perceive context. It can’t always tell whether a user is asking out of curiosity or out of desperation.
That’s why the most important part of any AI pastoral tool is often invisible: the guardrails. What happens when someone asks for something dangerous? What happens when someone expresses suicidal thoughts? What happens when someone requests advice that conflicts with medical or legal realities? What happens when a user is vulnerable and might interpret the AI’s words as authoritative in a way that bypasses human judgment?
In many AI deployments, these questions are addressed through policy layers, escalation pathways, and content restrictions. In a faith setting, the stakes are heightened because religious language can carry a sense of certainty. A comforting verse can become harmful if it delays urgent help. A “prayer” can become a substitute for therapy. A confident explanation can become a theological trap if it’s wrong or incomplete.
Reports about Lester’s twin suggest that the system is intended to support the church’s mission rather than operate independently. That implies a design philosophy that treats the AI as part of a larger ecosystem—one that includes human oversight, referral mechanisms, and a clear understanding of what the AI can and cannot do.
Even with safeguards, there is a broader cultural shift happening here—one that goes beyond any single pastor or church.
For decades, religious communities have relied on physical presence and scheduled programming: services, Bible studies, counseling appointments, youth groups, and volunteer rotations. Those structures create rhythm, but they also create friction. People who are traveling, working late, caring for family members, or simply hesitant to walk into a building can fall through the cracks. The AI twin, by contrast, offers a form of “always-on” access. It meets people where they already are: on their phones, in their private moments, when they’re alone with their thoughts.
That can be profoundly beneficial. Many people don’t want to wait days for reassurance. They don’t want to search for the right person to message. They want a response now—not because they’re impatient, but because their emotional state is immediate. In that sense, the twin functions like a digital front door to spiritual conversation.
But always-on access also changes expectations. When guidance is available instantly, users may begin to expect instant answers to everything. That can flatten the spiritual process, which often involves patience, reflection, and growth over time. Faith is not only about receiving information; it’s about learning to live differently. If the AI provides comfort without encouraging transformation, it risks becoming a vending machine for reassurance.
The unique opportunity—and responsibility—of a pastoral AI is to avoid that trap. A good system doesn’t just respond; it guides. It can ask follow-up questions that help a person articulate what they’re feeling. It can encourage journaling, suggest practical steps, and invite users to join real-world community activities. It can also remind users that spiritual care is not a transaction.
In other words, the twin’s value may depend on whether it helps people move from conversation to community.
There’s another angle that deserves attention: the twin as a model for how institutions might scale care without losing their identity.
Churches and nonprofits have long struggled with capacity. Volunteers burn out. Staff members are stretched thin. Even well-funded organizations can’t multiply hours in a day. AI offers a way to handle repetitive questions, provide basic guidance, and offer immediate support—freeing humans to focus on deeper, more complex needs.
Yet scaling care is not the same as scaling compassion. Compassion requires presence. It requires the willingness to sit with uncertainty and to respond to a person as a person. AI can simulate empathy through language, but it can’t fully replicate the moral commitment that comes from shared life.
That’s why the best implementations of AI in community settings tend to treat the technology as a supplement, not a replacement. The AI handles the first layer—questions, scripture references, prayer prompts, and general encouragement—while humans handle the second layer: discernment, crisis intervention, and relationship-building.
If Lester’s twin is indeed growing in daily use, it suggests that many people are finding the first layer helpful. The real test will be whether the system consistently routes users toward the second layer when needed.
There is also the question of how such a tool affects the pastor himself. Creating an AI twin can be emotionally strange. It means your words can travel without you. It means someone else can ask for “you” at 2 a.m., and the system will answer in your voice. That can feel empowering—like extending your ministry beyond your limits. It can also feel unsettling, like watching your identity become a product.
Pastors are used to being public figures, but an AI twin introduces a new kind of intimacy. It can talk to people privately, repeatedly, and without the social boundaries that exist in a church building. That raises questions about consent and representation: did the pastor fully anticipate how his likeness and style would be used? Did the church communicate clearly to users that they were interacting with an AI? Are users told how the system works and what it can do?
Transparency is not just an ethical requirement; it’s a trust requirement. Without it, users may feel deceived. With it, they can engage with the tool more responsibly.
The story also fits into a larger trend: faith leaders experimenting with AI not only for outreach, but for discipleship. Many churches are exploring chatbots for Bible study, automated reminders for events, and digital platforms for prayer requests. What Lester’s project adds is a more personal dimension—a conversational “presence” that resembles a specific pastor rather than a generic assistant.
That personalization is powerful. People