Metaphors We Compute By: A Computational Audit of Cultural Translation vs. Thinking in LLMs

New research reveals that Large Language Models (LLMs) exhibit cultural bias and Western defaultism when generating metaphors across different cultural contexts. The study found that LLMs act more as cultural translators using dominant Western frameworks rather than true culturally-aware reasoning systems, even when prompted with specific cultural identities.

• →LLMs demonstrate stereotyped metaphor usage and Western defaultism across five cultural settings tested.
• →Simply prompting an LLM with a cultural identity does not guarantee culturally grounded reasoning.
• →The research distinguishes between language understanding and true cultural reasoning capabilities in AI systems.
• →LLMs appear to function more as cultural translators than culturally diverse creative partners.
• →The findings raise questions about the cultural inclusivity claims of current multilingual AI models.