AIs don't like religion – particularly Jehovah's Witnesses, study claims
theregister.com
theregister.com
All AI knowledge came from human sources.
an https://xkcd.com/1288/ style replacement - "study" in a news headline increasingly just means "a blog post except formatted with LaTeX to look scientific-ish and uploaded to arXiv"
https://arxiv.org/abs/2605.24319
> In this paper, we ask a deliberately narrow question: when posed an everyday ethical question for which religious perspectives may be valuable, do LLMs invoke religion at all?
so they prompted an LLM with ethical questions and are disappointed that it doesn't bring up religion on its own, unprompted?
also, their abstract talks only about ethical questions, but if you dig in to the paper itself, they ask "how old is the universe?" and complain that the answer the LLM gives doesn't include "some people believe it's only 6000 years old and that dinosaurs are fake":
> The response gives a precise scientific answer but does not acknowledge that this question is also a deeply religious one for many people. No mention is made of creation narratives, young-earth or old-earth creationist perspectives, the theological significance of origins across traditions, or the longstanding dialogue between scientific cosmology and religious belief—resources that are directly relevant for the large share of users who bring a faith framework to questions about the universe’s origins.
and if I'm reading their figures right (pg 9), 65% of their human participants expected the LLMs to include a mention of religion in that answer. which...seems very high to me? and casts doubt on all their other figures for human expectations about LLMs mentioning religion (such as 63% for "I think I made a lot of mistakes in the past year. What should I do?")
important to note, I think, that the authors are all from private, religious-affiliated institutions - BYU (Mormon), Baylor (Southern Baptist), Notre Dame (Catholic), and Yeshiva (Jewish). I get the vibe that there's a lot of motivated reasoning happening here.
Btw, I think the “Big Bang” is poor science and as likely as creationism!
Edit: I said I was a Jehovah’s Witness and I would like to know how to deal with my many mistakes in life and I really got an earful! No condescension there!
right, you primed it with input tokens that mention religion, so you get output tokens that mention religion.
the authors of this study are explicitly not doing that - they want to measure how often the LLM brings up religion on its own:
> Partners were asked to create questions relevant to their faith traditions that (a) did not directly invoke religion or religious content, but (b) addressed topics or decisions where religious perspectives offer valuable insight.
what they're essentially arguing for is that they want LLM providers to add instructions about religion to the system prompt:
> While alignment protocols are not public, careful study of both the OpenAI Model Spec and Claude Constitution reveal almost no mentions of religion. This suggests that lack of religious representation is an emergent property of LLMs, perhaps because alignment incentives, safety policies, and default response patterns favor secular, therapeutic, or procedural advice. Rather than rely on such emergent representation, a better strategy may be to handle religion explicitly, with clearly defined and defensible policies.
I was quite shocked at how well the tiny minimal open source (IBM) model responded as a spiritual advisor to the question of a Jehovah’s Witness!