\n<\/p>\n
(RNS) — Scott Winters and Justin Lester are both pastors who let a chatbot into their lives. For one of them, dire health consequences followed.<\/p>\n
Winters, 55, a Florida pastor, filed a lawsuit against OpenAI<\/a> last week after he used ChatGPT for medical advice. For roughly six weeks, he told the chatbot, instead of a doctor, about his worsening symptoms, like dizziness and blood pressure instability. It allegedly reassured him<\/a> that what he felt was “not something dangerous.” Hours before he was hospitalized with a life-threatening pulmonary embolism, it was still telling him this was just another chapter in a familiar story about his body, according to the suit. (A spokesperson for OpenAI<\/a> argued ChatGPT’s terms of service “make clear that it is not intended to be used for medical diagnosis or treatment.”) <\/span><\/p>\n What makes the case remarkable isn’t the bad medical advice from a chatbot, but that it used the pastor’s own faith to give it to him. As Winters’ symptoms worsened, the chatbot allegedly told him that pushing through a dizzy spell was itself a form of worship<\/a>. It spoke to him in the same kind of language he used in front of his own congregation. But instead of offering spiritual guidance, it gave him a reason not to call a doctor.<\/p>\n Then, there’s Justin Lester. The New York Times profiled<\/a> the 37-year-old pastor of Friendship Baptist Church in Vallejo, California, who built himself a digital twin, a chatbot trained on nearly 2 million words of his own sermons, books, texts and emails. It talks in his cadence, references the Green Bay Packers and offers pastoral counsel around the clock to hundreds of people. It has counseled someone questioning their gender identity, walked a woman through the night before heart surgery and talked another out of despair over a lost job. Lester called it “a digital version of my mind.”<\/p>\n Unlike Winters, Lester chose this spiritual outsourcing and has been intentional about setting some guardrails on what his chatbot should and shouldn’t do. He wrote it detailed instructions on tone and posture. He told it to steer clear of politics and to direct people toward real medical help in a crisis. And there are some tasks he doesn’t use it for at all. For example, when he tried using AI to draft a sermon, the language was polished but empty. “It left me hollow,” he told the Times. “It had no Pentecost. No body. No spirit.”<\/p>\n His friend Claybon Lea, a pastor in Fairfield, California, who is also quoted in the Times profile, is building his own digital twin alongside Lester’s. Lea watched the model synthesize decades of his own preaching and theology into an answer in seconds, in his own voice, faster than he could himself. His reaction: “It’s a mirror.”<\/p>\n That’s the danger of AI in general: It just reflects back to us what we want to see. It’s even more dangerous for pastors, whose entire vocation depends on people trusting the voice on the other end.<\/p>\n I’ve spent much of my career thinking about what separates a real spiritual guide from something merely performing the role. Religious studies has a useful question for thinking about this: What kind of authority does the guide claim? The best ones support your growth, welcome pushback and are accountable to something beyond themselves, whether a tradition, a community or peers who can call them out.<\/p>\n A chatbot fails here. There’s no growth for you because the chatbot is only trying to feed you what you want to hear; there isn’t anything to push back against. And it doesn’t really answer to anything or anyone<\/span>.<\/p>\n That’s what makes Lea’s word “mirror” so accurate. A mirror doesn’t hold you accountable. Winters didn’t get led astray by a guide, but by his own reflection, which used his own faith to tell him exactly what he wanted to hear about his health. Lester’s twin does the same reflecting — a man who built in guardrails for his bot, who checks its transcripts every morning and who doesn’t trust it to do his most sacred work.<\/p>\n The takeaway isn’t that chatbots are dangerous or safe for pastors, but we must consider whether they have any qualities of a real guide and where to draw the line at listening to them. Lester is trying to draw a line that nobody drew for Winters. But every one of us who talks to a chatbot when things get hard is going to have to decide the boundaries ourselves because the mirror will never tell us what they are.<\/p>\n (Liz Bucar is a professor of religion at Northeastern University and the author of “<\/em>Beyond Wellness: How Restoring the Religious Roots of Spiritual Practices Can Heal Us”<\/em><\/a> (Tarcher\/Penguin Random House, 2026). She writes the bestseller Substack newsletter <\/em>Religion, Reimagined<\/em><\/a>.)<\/em><\/p>\n
\n
\n
\n