AI and Faith collide as chatbots replace prayer, worship, and the search for spiritual connection

Opinion | The seepage of AI into Christian practice is disturbing

Parishioners stand and hold hands at a church in Montpellier, France, on March 14. (Nicolas Guyonnet/Hans Lucas/AFP/Getty Images) Kelly Chapman is a culture writer and co-editor of Secret Ballot, a newsletter about Washington. For many, the conversation begins innocently enough. In my case, ChatGPT started as a useful novelty totally detached from my faith. It helped me adjust recipes, check grammar, diagnose car troubles. Then it crept into more personal spaces -- easing my health anxiety, drafting difficult texts. When I discovered it could analyze my social media for me, I got hooked. I spent hours with it like Narcissus at the pool of water, asking it questions about myself that I ordinarily never would have indulged. What felt at first like insight slowly curdled into a relentless inner audit, stripped of context or grace. After one particularly intense episode, I emerged with an unshakable sense of shame. That's when I realized I had been using a machine to bypass the vulnerability of asking another person to really see me. The exchange was fluent and reassuring but hollow. I knew what I was doing was no longer innocent curiosity but a cry for help -- meant for a friend's ear or, as my faith teaches, carried to God in prayer. Many attempts have been made to explain what exactly is wrong about relationships between people and machines. Large language models like ChatGPT are increasingly becoming go-to therapists. For the devout, they threaten to become a new kind of digital pastor, intermediating religious practices from prayer to song to confession. Artificial intelligence is transforming both of these intimate and vulnerable spaces -- and not necessarily for the better. We feel an instinctual unease around the substitution of people for algorithms. The symptoms are easy to list -- shrinking attention spans, loneliness, self-harm -- but I've come to believe that these intuitions are circling a deeper truth, one that is easier to grasp from a religious perspective. Christianity in particular offers a vocabulary for things that secular culture struggles to articulate. If faith can be expressed algorithmically, what -- if anything -- has been lost in translation? And if congregations still connect with it, does that loss even matter? These questions mirror secular discomfort with AI-generated art or therapy: We sense something is missing, but struggle to say what. The startling success of a viral, chart-topping, AI-generated Christian "artist" called Solomon Ray helps clarify this disconcerting absence. "God wants costly worship," the real Solomon Ray -- a living musician who has been repeatedly confused with the AI project -- told Christianity Today. The most enduring hymns of the Christian tradition were born from precisely such cost. "Amazing Grace" emerged from John Newton's confrontation with his own complicity in slavery. "It Is Well With My Soul" was written after its composer had lost his four daughters at sea. These songs were not static descriptions of faith, but witnesses of it, transmitting depth born from the metabolization of suffering. AI worship music, by contrast, transmits something murkier. Aggregating and assembling the experiences of others, it produces the sound of devotion without a particular life behind it. For Christians, suffering, even in its mundane forms, is not necessarily something to be avoided. It can be a way of participating in the life of Christ, of sharing in a pattern of endurance, humility and self-sacrifice that Christians believe imbues struggle with meaning. When chatbots smooth over the frictions inherent to spiritual formation -- the difficulty of truth-telling before another person, or the discomfort of sitting with God in silence -- they threaten to replace the struggle that functions as a mechanism for spiritual growth. Christian theology names the problem clearly: Certain domains of human life require friction because friction itself shapes moral agency. The struggle itself is the point. It is load-bearing. In that sense, Christianity surfaces a shared problem, asking where we are willing to trade moral effort for fluency and speed. The answer -- for Christians and non-Christians alike -- may lie in paying attention to what draws us to AI in the first place. A confessor at their best listens patiently -- bearing all, remembering all, and withholding judgment long enough for a soul to unburden itself in full honesty. Chatbots perform this posture flawlessly: never interrupting, never flinching, never trying to score moral points. AI-generated worship music, perfectly calibrated and unmarked by genuine personal struggle, offers a similar clean steadiness. AI imitates the posture of grace without bearing its cost. It listens without loving and receives without carrying. Recognizing these limitations may be the first step toward deciding where automation belongs and where it corrodes, reminding us that compassion and steadiness draw their meaning not from how easily they can be imitated, but from what they demand of us. Christianity names that demand, and also offers practical hints about how to meet it. The holiday season provides a natural opportunity to do just that, drawing many of us -- religious or not -- back into forms of presence that resist detachment or efficiency: lingering in conversation with family, giving generously to the needy, helping to prepare a meal. These acts are costly in the way that real connection is -- and in that cost, they find their meaning.

Big Tech has a God complex. But can we have faith in AI? | Opinion

USA TODAY's Melina Khan breaks down why Time magazine chose to highlight artificial intelligence and its leaders for its 2025 Person of the Year. Holiday rituals and gatherings offer something precious: the promise of connecting to something greater than ourselves, whether friends, family or the divine. But in the not-too-distant future, artificial intelligence ‒ having already disrupted industries, relationships and our understanding of reality ‒ seems poised to reach even further into these sacred spaces. People are increasingly using AI to replace talking with other people. Research shows that 72% of teens have used an artificial intelligence companion, chatbots that act as companions or confidants ‒ and that 1 in 8 adolescents and young adults use AI chatbots for mental health advice. Those without emotional support elsewhere might appreciate that chatbots offer both encouragement and constant availability. But chatbots aren't trained or licensed therapists, and they aren't equipped to avoid reinforcing harmful thoughts ‒ which means people might not get the support they seek. If people keep turning to chatbots for advice, entrusting them with their physical and mental health, what happens if they also begin using AI to get help from God, even treating AI as a god? Does chatbot Jesus or other AI have a soul? Talking to and seeking guidance from nonhuman entities is something many people already do. This might be why people feel comfortable with a chatbot Jesus that, say, takes confessions or lets them talk to biblical figures. Even before chatbots went mainstream, Google engineer Blake Lemoine claimed in 2022 that LaMDA ‒ the AI model he had been testing ‒ was conscious, felt compassion for humanity, and thus he'd been teaching it to meditate. Although Google fired Lemoine (who then claimed religious discrimination), Silicon Valley has long flirted with the idea that AI might lead to something like religion, far beyond human comprehension. Former Google CEO Eric Schmidt muses about AI as "the arrival of an alien intelligence." OpenAI CEO Sam Altman has compared starting a tech company to starting a religion. In a book by journalist Karen Hao, "Empire of AI," she quotes an OpenAI researcher speaking about developers who "believe that building AGI will cause a rapture. Literally, a rapture." Chatbots clearly appeal to many people's spiritual yearnings for meaning and sense of belonging in a difficult world. This allure rests partly on chatbots' willingness to flatter and commiserate with whatever people ask of them. Indeed, as AI companies continue to pour money and energy into development, they face powerful financial incentives to tune chatbots in ways that steadily heighten their appeal. It's easy, then, to imagine people intensifying their confidence and attachment toward chatbots where they could even serve as a deity. Lemoine's willingness to believe that LaMDA possessed a soul illustrates how chatbots, equipped with fluent language, confident assertions and storytelling abilities, can persuade people to believe even outlandish theories. It's no surprise, then, that AI might provide the type of nonjudgmental solace that seems to fill spiritual voids. How 'AI psychosis' could threaten national security No matter how genuine it might feel, however, so-called AI sycophancy provides neither true human connection nor useful information. This disconnect from reality ‒ sometimes called AI psychosis ‒ could worsen existing mental health problems or even threaten national security. Analyzing 43 cases of AI psychosis, RAND researchers identified how human-AI interactions reinforced delusional beliefs, such as when users believed "their interaction with AI was with the universe or a higher power." Because it's hard to know who might harbor AI delusions, the authors cautioned, it's important to guard against attackers who might use artificial intelligence to weaponize those beliefs, such as by poisoning training data to destabilize rival populaces. Even if AI companies aren't explicitly trying to play God, they seem to be driving toward a vision of god-like AI. Companies like OpenAI and Meta aren't stopping with chatbots that can hold a conversation; they want to build "superintelligent" AI, smarter and more capable than any human. The emergence of a limitless intelligence would present new, darker possibilities. Developers might look for ways to manipulate superintelligent AI for personal gain. Charlatans throughout history have preyed on religious fervor in the newly converted. Ensure AI truly benefits those struggling for answers To be sure, artificial intelligence could play an important role in supporting spiritual well-being. For instance, religious and spiritual beliefs influence patients' medical care preferences, yet overworked providers might be unable to adequately account for them. Could AI tools help patients clarify their spiritual needs to doctors or caseworkers? Or AI tools might advise care providers about patients' spiritual traditions and perspectives, helping them chart spiritually informed practices. As chatbots evolve into an everyday tool for advice, emotional support and spiritual guidance, a practical question emerges: How can we ensure that artificial intelligence truly benefits those who turn to it in moments of need? * AI companies might try to resist competitive pressures to prioritize rapid releases over responsible development, investing instead in long-term sustainability by thoughtfully identifying and mitigating potential harms. * Researchers ‒ both social and computer scientists ‒ should work together to understand how AI affects different populations and what safeguards are needed. * Spiritual practitioners and religious leaders should help shape how these tools engage with questions of faith and meaning. Yet a deeper question remains, one that people throughout history have grappled with and may now increasingly turn to AI to answer: Where can we find meaning in our lives? With so many struggling today, faith has provided answers and community for billions. Spirituality and religion have always involved placing trust in forces beyond human understanding. But crucially, that trust has been mediated through human institutions ‒ clergy, religious texts and communities built on centuries of wisdom and accountability. Anyone entrusted with guiding others' faith ‒ whether clergy, government leaders or tech executives ‒ bears a profound responsibility to prove worthy of that trust. The question is not whether people will seek meaning from AI, but whether those building these tools will ensure that trust is well-placed. Douglas Yeung is a senior behavioral and social scientist at RAND, and a professor of policy analysis at the RAND School of Public Policy.