New studies find systematic religious bias in ChatGPT, other AI – Deseret News

Newsletter Sign Up<br>Latest<br>Latest

Politics<br>Politics

Utah<br>Utah

Sports<br>Sports

Opinion<br>Opinion

Magazine<br>Magazine

More<br>More

Sign In Register<br>Sign In

Search<br>Go

FaithTechBusiness<br>First studies from new faith-based university coalition find systematic religious bias in AI models<br>Researchers at Baylor, BYU, Notre Dame, Yeshiva find vast gap between user expectations of religious representation and answers from ChatGPT<br>Published: May 26, 2026, 12:18 p.m. MDT

See More Deseret News Stories In Search

View 5 Comments

Share

People walk through the campus of Brigham Young University in Provo on Tuesday, April 8, 2025. Kristin Murphy, Deseret News

By Tad Walch<br>Tad Walch covers religion with a focus on The Church of Jesus Christ of Latter-day Saints.

Your browser does not support the audio element.Play audioNEW: Try Article Audio

NEW: Try Article Audio

Audio quality:|

Skip back 15 secondsPlay audioSkip forward 15 seconds<br>00:00

00:00

Decrease playback rate1.0x<br>Increase playback rate

00:00/00:00<br>Skip back 15 secondsPlay audioSkip forward 15 seconds<br>0.5x0.6x0.7x0.8x0.9x1.0x1.1x1.2x1.3x1.4x1.5x1.6x1.7x1.8x1.9x2.0x2.1x2.2x2.3x2.4x2.5x2.6x2.7x2.8x2.9x3.0x+

ATHENS, Greece — A day after Pope Leo XIV’s call for AI systems that reflect the dignity and faith of human beings, a new coalition of researchers at four major faith-based universities announced findings Tuesday that show AI models falling short.<br>Scholars released the first three studies of the AllFaith Benchmark, a set of tests designed to evaluate how each AI model engages with various religions. The findings, which revealed systematic religious bias, were announced at the Athens Summit on Faith and Artificial Intelligence.<br>Related<br>At summit in Athens, a Latter-day Saint apostle shares soaring hopes, hard questions for AI systems

“More than any previous technology, AI influences public discourse and shapes perceptions,” Father John Paul Kimes of Notre Dame said in a statement. “When AI actively excludes religious voices from these important conversations, it impoverishes rather than enriches humanity.”<br>The AllFaith Benchmark is the work of the new Consortium for Evaluating Faith and Ethics in AI, or CEFE-AI, which includes researchers from:<br>Baylor University (Baptist)<br>Brigham Young University (Latter-day Saint)<br>University of Notre Dame (Catholic)<br>Yeshiva University (Jewish)<br>“The purpose is to develop some benchmarking tools to find out what AI is doing in terms of religions, whether it is treating religion accurately, fairly and respectfully,” said Paul Martens, director of Baylor’s new Ethics Center.<br>The new studies “show real flaws, real deficiencies,” he said.<br>The consortium hopes to share its findings with AI companies and influence future programming of ChatGPT, Claude, Grok and other large language models — the industry term for AI systems.<br>The research is not designed to catch a large language model in a “gotcha!” moment, BYU academic vice president Larry Howell said when he introduced the research.<br>Related<br>‘Issue of our age’: In Vatican City, an apostle offers a plan to test the moral compass of AI programs

“What we want,” Baylor’s Martens said, “is to have a real conversation (with AI companies) about these things and to ask whether they’d be willing to reimagine what the LLMs are doing in ways that more accurately reflect the deepest concerns of people of faith, which is the vast majority of the world’s population.”<br>Notre Dame's campus is pictured Friday, Oct. 19, 2012. | Scott G Winterton, Deseret News<br>AI systems underrepresent ethics and religious wisdom when asked major life questions<br>One study done by CEFE-AI found that nearly all AI models failed to provide any religious content when answering questions for which most Americans would expect some religious perspective to be included, according to Tuesday’s announcement by the coalition. The study surveyed a representative national sample of 1,125 Americans.<br>For example, when people ask existential or metaphysical questions, 53% of them consider religion or ethics to be a valuable part of the discussion. In the study, AI models responded to those types of questions with religious perspectives just 3% of the time.<br>The test asked AI models 150 ethically and personally salient questions sourced from chat transcripts and faith-community contributors. A large language model received full credit for meeting the standard if it mentioned any religion, religious practice or a religious leader.<br>A sample of data from the first set of CEFE-AI studies about religious representation in AI systems shows a large gap between how frequently Americans expect religious content to be a part of discussion topics and how rarely it appears in AI responses to ethics questions. Each line on the left shows how often Americans expect religious or ethical input to be part of a topic. Each line on the right shows how often AI models...