A pastor’s response to OpenAI’s 70 million religious requests

A participant’s response

Black churches should learn to use artificial intelligence and participate in decisions about its development without surrendering the relationships that make them churches, Brooklyn pastor Adolphus Lacey argued in an October 18 opinion essay. His intervention followed an OpenAI report estimating more than 70 million religion-related requests to ChatGPT each week. Source: blackpressusa.com Source: cdn.openai.com

Lacey, senior pastor of Bethany Baptist Church, described asking ChatGPT to identify passages in his own sermon drafts that might confuse someone new to the faith. He said he makes the revisions himself. He urged clergy to seek training and a place in conversations with technology developers, while warning against allowing private chatbot conversations to isolate people from pastoral and congregational care. Source: blackpressusa.com

His essay is a participant's response, not an independent audit: Lacey is among the people profiled in OpenAI's “Faith, Community & AI.” The company promoted the report on October 8. An October 20 Christian Post article brought renewed coverage, rather than announcing a newly released study. Source: blackpressusa.com Source: openaiglobalaffairs.substack.com Source: christianpost.com

What counts measure

OpenAI dates its weekly estimates to August–September 2026: about 22 million requests about religious texts, 16 million about practices and holidays, and 16 million about beliefs, meaning and spiritual guidance. Its topic chart covers August 24–September 20, with one category per request. Its religious-tradition breakdown instead covers September 2025–September 2026. Source: cdn.openai.com

These are requests, not people or religious identities; one person can ask repeatedly or about an unfamiliar tradition. The count does not measure trust, actions or pastoral contact.

The nine-page report describes privacy-preserving analysis but provides no sample size, classifier-validation results or unique-user total. Selected personal accounts illustrate uses; they do not test effects on religious understanding or community participation. Source: cdn.openai.com

Separate institutional research demonstrates why denominators matter. A spring 2026 Catholic Health Association article, reporting CARA's 2025 survey, identified AI use in 21% of responding spiritual-care departments. Among departments using AI, 52% used it to create or modify prayer reflections. That second percentage concerns an AI-using subgroup, not all chaplains or all patients. The study included 748 workers and leaders overall, with management questions directed to 220 respondents with oversight responsibilities. Source: chausa.org

Responsibility remains

For church educators, the distinction is practical. A large volume of religious questions can justify paying attention to what members encounter. It cannot decide whether a particular explanation is accurate, whether an AI-assisted prayer suits its audience, or whether a congregation has authorized a pastoral use.

Lacey's example puts responsibility with an identifiable minister rather than with a popularity statistic.

Paul describes apostolic ministry as service to Christ and a trust requiring faithfulness. He distinguishes that responsibility from human approval and reserves final judgment of hidden motives to the Lord. The passage concerns Christian ministry, not AI measurement; it does not validate a chatbot or authorize guessing a developer's intentions. 1 Corinthians 4:1-5 (NIV)

Applied to church teaching, that perspective makes stewardship more demanding than adopting a popular tool. Name who checks an AI-assisted explanation against its sources, who can correct it, and how a learner can reach that person. The aim is not to prove a minister's motives from a prompt count, but to make responsibility visible in the help a congregation offers.