When Your Pastor Is an AI: A Bay Area Church's Digital Double Raises Old Questions in New Code
Justin Lester built an AI version of himself to answer congregants around the clock — and more people are talking to it every day
By Ada "Peer-Review" Sparks
Justin Lester couldn't be everywhere his congregation needed him — so he built a version of himself that could. According to reporting in The New York Times, the Bay Area pastor created an AI "twin" trained to discuss faith and answer questions about God, available to churchgoers at any hour.
The Times reports that usage is climbing: more people are turning to the AI duplicate every day to talk through matters of belief that once required a phone call, an office visit, or a Sunday-morning conversation. The appeal is obvious — pastoral care doesn't keep business hours, and a chatbot never sleeps.
What isn't yet clear, and what the available reporting doesn't establish, is how the AI twin was built, what data or model underlies it, how its answers are vetted for theological accuracy, or what happens when it gets something wrong. Spiritual counsel is not customer service; a bad answer about grief, doubt, or crisis carries different stakes than a wrong restaurant recommendation. No outside theologian, ethicist, or AI safety researcher is quoted assessing the tool in the reporting we have.
This fits a broader pattern this desk watches closely: institutions — churches, clinics, schools — adopting generative AI to close real gaps in access, often faster than anyone builds guardrails to match. The technology is real and the demand is real. Whether an algorithm should be dispensing spiritual guidance, and who is accountable when it errs, remains an open question this single account doesn't answer.
We'd want to know: what model powers the twin, who reviews its output, and whether the church tracks cases where it gave counsel a human pastor would not have.
Extraordinary claims. Ordinary evidence? Then no.
— Compiled from reporting by The New York Times.
The American Times' desks are written under standing pen names; the reporting under every byline meets the paper's sourcing standards. See "About Our Bylines."