The Narrative Gap: Can LLMs Help Us Navigate Diverse Narratives Across Languages?
arXiv cs.CL · 10 Nov 2026, 03:31 AEST
Large language models (LLMs) are increasingly mediating how people seek and consume information across different languages. Current models lack the necessary capabilities to reliably navigate diverse narratives, which can lead to significant gaps in users' exposure to competing accounts.
Key facts
- Supporting diverse information seeking requires models to help users navigate competing narratives across languages and information streams.
- We define six capabilities necessary for diverse multilingual information seeking: three foundational capabilities for narrative inclusion, narrative exclusion, and missing narrative identification; and three higher-level capabilities for presenting diverse narratives, preserving narrative coverage under confirmatory queries, and maintaining information parity across languages.
- Evaluating 14 LLMs across 24 query languages, we find that current models lack the capabilities needed to reliably navigate diverse narratives, limiting users' control over the perspectives they receive and leaving substantial gaps in exposure to competing accounts even when diversity is explicitly requested.
Read aloud
Large language models L L M s are increasingly mediating how people seek and consume information across different languages. Current models lack the necessary capabilities to reliably navigate diverse narratives, which can lead to significant gaps in users' exposure to competing accounts. Supporting diverse information seeking requires models to help users navigate competing narratives across languages and information streams. We define six capabilities necessary for diverse multilingual information seeking: three foundational capabilities for narrative inclusion, narrative exclusion, and missing narrative identification; and three higher-level capabilities for presenting diverse narratives, preserving narrative coverage under confirmatory queries, and maintaining information parity across languages. Evaluating fourteen L L M s across twenty four query languages, we find that current models lack the capabilities needed to reliably navigate diverse narratives, limiting users' control over the perspectives they receive and leaving substantial gaps in exposure to competing accounts even when diversity is explicitly requested.
Sources in this cluster
- The Narrative Gap: Can LLMs Help Us Navigate Diverse Narratives Across Languages?
arXiv cs.CL · 10 Nov 2026, 03:31 AEST