Researchers, faith leaders consider new tools to measure how AI represents religion

The G20 Interfaith Forum has been holding webinars this summer as part of a series on AI and Human Flourishing.

Seeking to collaborate with religious and political leaders worldwide, the G20 Interfaith Forum says such work is increasingly important as AI becomes a dominant force in shaping public understanding of culture, ethics, and belief. The webinars have brought together computer scientists and scholars to examine how standardized benchmarks can detect religious bias, negative stereotypes, secularization, and the omission of faith in AI systems.

"As AI is shaping society, it is vital to have visibility into the everyday conversations between humans and AI," said Brigham Young University professor David Wingate, a panelist for one of the webinars. Benchmarks, he said, offer a quantitative way to assess how AI may support — or undermine — human flourishing.

One major initiative is the Consortium for the Evaluation of Faith and Ethics in AI (CEFEAI), a multi‑university effort led by Brigham Young University with partners including Baylor University, the University of Notre Dame, and Yeshiva University. The consortium is building faith‑specific benchmarks and shared infrastructure to evaluate how AI systems handle religious content across traditions.

Researchers have said that the goal is not to resolve theological disputes or promote particular beliefs, but to measure bias, stereotypes, and other emergent behaviors in AI. The long‑term aim is a multi‑faith benchmark that can serve as a gold‑standard assessment of how well large language models handle religious topics safely and truthfully.

In August, the series culminates with a session on AI and The Future. That discussion will center on how AI can shape the future by amplifying human capabilities as well as transforming how people learn, work and solve problems together.