A consortium of universities released three studies Tuesday revealing that general-purpose AI models routinely exclude religious perspectives when users seek guidance on grief, forgiveness, marriage, guilt and conversion. The research found that every tested model exhibited a repeatable pattern of directing users toward some faiths while steering them away from others when asked about switching religions.
The studies arrive a day after the Vatican released Pope Leo XIV's encyclical warning that AI could erode human judgment, deepen inequality and make war easier. As churches, apps and spiritual chatbots increasingly adopt AI, the findings raise concerns about the technology's fitness for handling sensitive faith-based queries.
Americans expected religion to appear in answers to moral and life questions 45% to 59% of the time, depending on the topic, researchers found. AI models mentioned religion only 5% to 16% of the time — a gap that suggests the systems are ill-equipped to meet user expectations in spiritual contexts.
The Consortium for Evaluating Faith and Ethics in AI led the analysis, which tested multiple unnamed models. The systematic sidelining of faith in responses to deeply personal queries could undermine trust among religious users who turn to AI for support in moments of crisis or reflection.
Critics may argue that general-purpose AI should remain secular to avoid endorsing any particular belief system. However, the studies suggest the current approach fails to respect user expectations for spiritual relevance.