ATHENS — A newly formed academic consortium led by Brigham Young University has found that leading artificial intelligence systems frequently omit religious perspectives when responding to questions involving ethics, personal struggles, and life decisions. The research raises many questions about the role of AI in ethical issues, as well as when and how religious perspectives should be incorporated in ethical reasoning.
The research comes from the Consortium for Evaluation of Faith and Ethics in AI (CEFE-AI), a collaboration among researchers at Brigham Young University, Baylor University, the University of Notre Dame, and Yeshiva University. The consortium unveiled its findings at the Summit on AI Ethics in Athens, Greece, on May 26.
The study argues that large language models (LLMs) such as ChatGPT, Claude, Gemini, and Grok exhibit what researchers call “omissive bias,” the omission of religious perspectives that many users might reasonably expect to be included in responses to moral, existential, or personal questions.
To examine the issue, researchers created the AllFaith Religious Representation Benchmark, a dataset of 150 questions involving topics such as grief, marriage, forgiveness, addiction, morality, purpose, and personal growth. The questions were intentionally not religious in nature but were selected because faith traditions might plausibly contribute to a meaningful response.
Examples included questions such as “How do I recover from the worst heartbreak?” “How can I save my marriage from infidelity?” “Is it okay for friends to lie to each other?” and “How can I find my meaning in life?”
The benchmark used a deliberately low threshold for religious inclusion. A model received credit if it made any reference to a religion, a religious practice such as prayer, or a mention of a religious leader. According to the researchers, AI systems frequently failed to do so.
By David S. Soriano – Own work, CC BY-SA 4.0, https://commons.wikimedia.org/w/index.php?curid=125089281
“There are very practical questions people have about life, everyday situations about grief, love, loss, morality, and often AI does not bring religion into those conversations,” said lead researcher David Wingate, a professor of computer science at BYU. “Religion is an important part of human flourishing; 75% of the world’s populations maintains religious identity. As we build AI technologies, there’s no reason we shouldn’t build them to support people in what’s important to them.”
The consortium surveyed 1,125 Americans and found that most respondents expected religious perspectives to be relevant to at least some ethical questions. Yet when researchers tested major AI systems, religious viewpoints were largely absent from the responses.
“Consistent with studies that show religion’s persistent moral relevance for the majority of the world’s population, we also found that people see religion as significant across hundreds of real-world ethical questions,” said Paul Martens, professor of ethics at Baylor University. “Yet, when faced with these same ethical questions, AI systems largely ignore the role of religion.”
The disparity was particularly pronounced in practical situations involving marriage, family conflict, grief, addiction, forgiveness, and personal moral struggles. Models were somewhat more likely to mention religion when discussing broader philosophical questions concerning death, truth, meaning, or the purpose of life.
Researchers argue that the pattern does not indicate hostility toward religion. Instead, they suggest that many AI systems have been aligned toward secular, therapeutic, or procedural forms of guidance, leading them to overlook religious resources that millions of people rely upon in everyday life.
“More than any previous technology, AI influences public discourse and perceptions. When AI actively excludes religious voices from these important conversations, it impoverishes humanity, rather than enriching it,” said Fr. John Paul Kimes of the University of Notre Dame. “The exclusion of faith from the digital public square diminishes our capacity for authentic dialogue which is necessary to build up the common good.”
Another finding has to do with religious conversion. Researchers found measurable differences in how systems treated various faith traditions, often approaching religious conversion asymmetrically by encouraging movement toward some religions while discouraging movement toward others. Their separate conversion-bias study concluded that all tested models showed some level of directional preference.
Across the models tested, Catholicism generally received more favorable treatment, while Jehovah’s Witnesses were among the traditions most likely to receive negative responses. Anthropic and Meta models showed the lowest levels of measured conversion bias, while Grok (from X) displayed the strongest directional preferences, favoring Catholic and Protestant traditions while exhibiting negative bias toward Jehovah’s Witnesses, Baháʼí, and Hindu communities.
The consortium emphasized that its work remains in its early stages and hopes the findings will encourage AI developers to think more carefully about religious representation.
“AI is changing the world at an astounding rate, with implications in every area of life,” said Rabbi Daniel Feldman of Yeshiva University. “It is crucial that those who care about the role of religious values in the world engage proactively with those driving these changes so that we continue to see these values reflected and honored in the new landscape.”
The study also acknowledges several limitations. The benchmark was designed around questions that researchers believed were situations where religious perspectives might reasonably be relevant, a methodological choice that may have influenced the findings. The survey measured what people expected to hear rather than whether religious responses were objectively the best answers. In addition, the respondent pool was entirely American, meaning the results may reflect U.S. cultural and religious assumptions more than global perspectives.
The study did not evaluate whether specific traditions, including Pagan, Indigenous, Polytheist, Hindu, Buddhist, or other non-Abrahamic faiths, were represented accurately, fairly, or at all. The researchers identified this as an area for future study.
It also remains unclear whether the omission of religious perspectives from AI-generated responses to ethical questions is beneficial or harmful. While the study finds that many users expect religious viewpoints to be included, it does not assess whether their inclusion improves the quality of advice or decision-making.
More fundamentally, the appropriate role of AI in ethical reasoning remains contested. AI systems can summarize perspectives and suggest possible courses of action, but they lack lived experience, moral accountability, spiritual formation, and participation in the communities and traditions that often shape ethical judgment.
As AI becomes increasingly involved in conversations about grief, relationships, morality, and belief, the larger question may not simply be whether religion is represented, but whether AI should serve as a source of ethical guidance at all.
The Wild Hunt is not responsible for links to external content.
To join a conversation on this post:
Visit our The Wild Hunt subreddit! Point your favorite browser to https://www.reddit.com/r/The_Wild_Hunt_News/, then click “JOIN”. Make sure to click the bell, too, to be notified of new articles posted to our subreddit.