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.
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.
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.
Many don’t have any issue with AI not bringing religion into conversation. Some actually prefer it. Even a subtle nudge toward one religion over another is a serious problem, and AI companies owe it to their users to fix it.
This article is shared at no charge for educational and informational purposes only.
Red Sky Alliance is a Cyber Threat Analysis and Intelligence Service organization. We provide indicators of compromise information (CTI) via a notification/Tier I analysis service (RedXray) or an analysis service (CTAC). For questions, comments or assistance, please contact the office directly at 1-844-492-7225, or feedback@redskyalliance.com
Weekly Cyber Intelligence Briefings:
- Reporting: https://www.redskyalliance.org/
- Website: https://www.redskyalliance.com/
- LinkedIn: https://www.linkedin.com/company/64265941
Weekly Cyber Intelligence Briefings:
REDSHORTS - Weekly Cyber Intelligence Briefings
Comments