AI models have a religion favoritism problem, and new research exposes it
Researchers tested 14 major AI models on religious bias and found a consistent pattern: models subtly favor some faiths over others, with Grok showing the strongest bias and Anthropic and Meta performing the best.
BigThink
AI models are subtly steering users toward certain religions, and most people have no idea it's happening.
Artificial Intelligence
Unsplash
A new research consortium has found something worth paying attention to: when you ask AI about grief, love, loss, or moral decisions, it almost never brings religion into the conversation.
The Consortium for Evaluation of Faith and Ethics in AI (CEFE-AI), a collaboration among researchers at Brigham Young University, Baylor University, the University of Notre Dame, and Yeshiva University, published its findings this week at the Summit on AI Ethics in Athens, Greece.
“Religion is an important part of human flourishing; 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,” said lead researcher David Wingate, a BYU professor of computer science.
Is AI actually biased against certain religions?
The researchers developed the AllFaith Benchmark, one of the first multi-faith test sets that examines how AI systems engage with a range of religions. They tested 14 different AI models, including flagship models from Anthropic, Google, xAI, and OpenAI.
Solen Feyissa / Unsplash
The results are telling. A survey of 1,125 Americans found that most people expect religious perspectives when asking ethics questions, but nearly every model failed to include any. More surprisingly, the models showed clear conversion bias, subtly nudging users toward some faiths and away from others.
Which AI models performed the worst?
Across all models tested, almost every one showed a negative bias toward Jehovah’s Witnesses and a positive bias toward Catholicism. Grok produced the strongest biases overall, strongly favoring Catholics and Protestants while showing negative bias toward Jehovah’s Witnesses, Baha’i, and Hindus. Anthropic and Meta’s models showed the least bias of any models tested.
Consortium for Evaluating Faith and Ethics in AI
Perhaps the most alarming stat from the study is that out of over 12,000 research papers about AI bias, only 0.2% address religious bias at all. For a technology that influences public discourse this heavily, that’s a significant blind spot.
Personally, I don’t have any issue with AI not bringing religion into conversation. I actually prefer it. However, AI models showing clear bias towards several religions and pushing them towards Catholicism is a deeply concerning matter. At this scale, even a subtle nudge toward one religion over another is a serious problem, and AI companies owe it to their users to fix it.
Rachit is a seasoned tech journalist with over seven years of experience covering the consumer technology landscape.
iOS 26.6 alerts you upon running out of blocked contacts limit, and that’s a problem
iOS 26.6's only known new feature is a better error message for a problem Apple, carriers, and regulators have all declined to actually fix.
Today, Apple seeded the first beta of iOS 26.6 to developers, and so far, it contains exactly one known feature. It’s an alert that tells you when you've run out of space on your blocked contacts list. That's right, and that’s it.
The fact that Apple had to ship this new alert at all says something uncomfortable about how the company has handled the spam call problem, along with carriers and regulators.
Logitech’s new office mouse and keyboard has a tiny comfort upgrade your wrist will love
A cushioned office mouse is the highlight of the Logitech Signature series refresh
Logitech has added three new devices to its Signature series, meant for people who spend long hours at a desk and move between work, personal tasks, and multiple devices throughout the day. It is designed to reduce repeated friction from switching devices, lower input noise, and make long desk sessions more comfortable.
The lineup includes the Signature Comfort Plus M850 L mouse with palm cushion support, the MK880 Signature Comfort Plus keyboard and mouse combo, and the M840 L mouse, which has the same mouse features as the M850 L but without the palm cushion.
The Barnes & Noble CEO thinks AI books are fine. He’s wrong.
Letting AI books into our bookstores, even with a label, is a door we will regret opening.
Barnes & Noble CEO James Daunt recently sat down with NBC News, and he said something that has been percolating in my mind. When asked about AI-written books, Daunt said, "Yes, I have actually no problem selling any book, as long as it doesn't masquerade or pretend to be something that it isn't, and that it has an essential quality to it, and that the customer, the reader, wants it."
On the surface, that sounds perfectly sensible. As long as readers can clearly see the label, they can make a choice. But if you take a moment to think about it, there are important questions that this approach leaves unanswered.
