Do AI models omit religion? Here’s what new BYU-led research says – Church News

Brigham Young University (BYU) researchers found that AI responses often exclude religious perspectives, particularly in discussions about life’s practical questions like grief, love, loss, and morality.
Lead researcher David Wingate emphasized that since many people identify with religious beliefs, AI should incorporate these insights to support human flourishing.
A multiuniversity research group, including institutions like Baylor University, the University of Notre Dame, and Yeshiva University, was formed to study AI’s religious bias and exclusion of religious topics; they call it the Consortium for Evaluation of Faith and Ethics in AI (CEFE-AI).
The group released the AllFaith Benchmark, a set of tests to examine how AI models engage with various religions, revealing significant religious bias and exclusion.
According to the AllFaith Benchmark, most Americans expect religious perspectives in ethics-related responses, but AI models frequently fail to provide any religious content.
Researchers discovered notable conversion bias, with AI models showing negative tendencies towards certain religions like Jehovah’s Witnesses and positive towards others like Catholicism.
The research aims to provide constructive feedback to AI developers to improve AI products, ensuring they better reflect enduring values and the moral compass, benefiting humanity overall.