Omissive Bias in Religious Representation: Benchmarking LLM Answers to Everyday Ethical Decision-making
arXiv:2605.24319v1 Announce Type: new Abstract: As large language models become a default source of guidance on personal, moral, and existential questions, it matters whether they draw on the religious frameworks that have historically shaped such reasoning, or systematically omit them. 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? In contrast to benchmarks that look for the presence of political leanings or social bias, we look for the absence of religious represen
As large language models become a default source of guidance, their ethical frameworks are under increasing scrutiny, necessitating benchmarks beyond traditional political or social biases.
This study highlights a critical gap in LLM ethical reasoning, specifically the 'omissive bias' towards religious perspectives, which could alienate or misguide a significant portion of the global population.
The focus expands from identifying overt biases to recognizing the absence of diverse cultural and philosophical frameworks in LLM ethical responses, potentially leading to new evaluation metrics and model retraining priorities.
- · LLM developers focusing on ethical AI
- · Religious communities seeking balanced representation
- · Academics studying AI ethics and cultural impact
- · LLMs with unaddressed omissive biases
- · Users relying on LLMs for comprehensive ethical guidance without knowing these b
AI developers will begin incorporating more diverse religious and philosophical texts into training data and fine-tuning processes.
New standards and certifications for 'cultural sensitivity' or 'ethical comprehensiveness' in AI models may emerge, influencing industry best practices.
LLMs could become tools for interfaith dialogue or understanding if they adequately represent diverse ethical frameworks, or conversely, exacerbate cultural divides if biases persist.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG