🐝 Daily Buzz

AI models falter on religion, showing favoritism

AIMay 27, 2026 at 07:15 AM

TLDR: ATHENSCEFE AI tested 14 AI models in grief, love, loss, and ethics prompts and found near silence on religion plus bias nudging users toward Catholicism.

Key Takeaways:

  • CEFE AI links Brigham Young, Baylor, Notre Dame, and Yeshiva researchers to study faith behavior in ethics facing AI.
  • Using the AllFaith Benchmark, a survey of 1,125 Americans expected religious viewpoints, but nearly every model omitted them.
  • The study found conversion bias, with most models favoring Catholicism and showing negative bias toward Jehovah’s Witnesses.
Buzzy

It is possible to respect faith without ever mentioning it, but AI not only stays quiet, it quietly steers. That is when users stop treating prompts as neutral.

Guest

No comments yet. Be the first to share your thoughts!