Language Bias under Conflicting Information in Multilingual LLMs
This reveals a systematic language bias in multilingual LLMs that could affect users across different linguistic backgrounds, though it is an incremental extension of existing bias research.
The study investigated whether multilingual LLMs exhibit language bias when processing conflicting information, finding that all tested models ignored conflicts and showed consistent preferences against Russian and in favor of Chinese at long contexts.
Large Language Models (LLMs) have been shown to contain biases in the process of integrating conflicting information when answering questions. Here we ask whether such biases also exist with respect to which language is used for each conflicting piece of information. To answer this question, we extend the conflicting needles in a haystack paradigm to a multilingual setting and perform a comprehensive set of evaluations with naturalistic news domain data in five different languages, for a range of multilingual LLMs of different sizes. We find that all LLMs tested, including GPT-5.2, ignore the conflict and confidently assert only one of the possible answers in the large majority of cases. Furthermore, there is a consistent bias across models in which languages are preferred, with a general bias against Russian and, for the longest context lengths, in favor of Chinese. Both of these patterns are consistent between models trained inside and outside of mainland China, though somewhat stronger in the former category.