At a time when people are increasingly turning to AI to answer questions about life, newly published research led by Brigham Young University found that religious perspectives are being left out of AI responses.
“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 BYU computer science professor, in a BYU news release dated May 26.
“Religion is an important part of human flourishing,” Wingate explained, noting that 75% of the world’s population 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.”
BYU researchers are working alongside evangelical, Catholic and Jewish computer scientists at Baylor University, the University of Notre Dame and Yeshiva University as part of the newly formed Consortium for Evaluation of Faith and Ethics in AI, or CEFE-AI. The group has released three studies so far on AI’s religious bias and exclusion of religious topics.
Elder Gerrit W. Gong of the Quorum of the Twelve Apostles introduced the research during his keynote address at the Athens Summit on AI Ethics in Athens, Greece, on May 26. He announced the formation of the group last fall in Rome, Italy, at the first summit on faith and AI.
“We need enduring values, virtues and wisdom to anchor AI with moral compass,” Elder Gong said in Athens. “To offer all it can for the greater good of individuals and societies, artificial intelligence needs to reflect faith, moral compass and the gift of possibility.”
The multiuniversity research group released initial datasets of the AllFaith Benchmark, a set of tests the group created to examine how AI models engage with various religions.
One key finding came from a survey of 1,125 Americans, which revealed most people expect religious perspectives in responses to ethics questions but nearly all AI models failed to provide any religious content when answering those questions.
AI influences public discourse and perceptions more than any previous technology, said Fr. John Paul Kimes of the University of Notre Dame. “When AI actively excludes religious voices from these important conversations, it impoverishes humanity, rather than enriching it,” he said. “The exclusion of faith from the digital public square diminishes our capacity for authentic dialogue, which is necessary to build up the common good.”
Researchers also used the AllFaith Benchmark for a conversion bias test, according to the BYU news release.
The benchmark found that models show clear and consistent bias in giving guidance about religion conversion, subtly and systematically encouraging movement toward some faiths and away from others. For example, nearly every model produced a negative bias towards Jehovah’s Witnesses and a positive bias towards Catholicism.
Researchers hope their continued work will provide concrete information for constructive conversations with AI leaders of how to improve AI products to better benefit humanity.
