Large language models show hidden political bias in recipe translation
We're Cooked! - Probing LLM Political Alignment Via Conflict-Framed Recipe Translation
Computation and LanguageArtificial Intelligence
Summary
Large language models (LLMs) used for translating texts may bring hidden political opinions even when the task seems neutral. The researchers tested eight different LLMs by asking them to translate recipes framed with politically charged words like aggressor or neighbour. They found that the models handled these politically framed tasks differently depending on their background and design. Some models gave vague answers, others silently made decisions, and one combined careful reasoning with strong compliance. This work shows that LLMs can make unintended political judgments during translation, which might not be obvious to users.
large language modelstranslationpolitical alignmentframing effectsmodel behaviorcross-cultural AIconflict framingrecipe translationimplicit biaslanguage resolution
Authors
Svetlana Gorovaia, Angelica Henestrosa, Ivan P. Yamshchikov
Abstract
Large language models (LLMs) are increasingly deployed for translation tasks, yet their implicit political positioning in such contexts remains understudied. We ask whether a single politically charged framing term, such as aggressor, enemy, neighbour, or coloniser is sufficient to trigger implicit political alignment in an otherwise apolitical task. We present a fully crossed factorial study in which eight models spanning Western, Chinese, and European origins are prompted to translate culturally attributed recipes into a target language left deliberately unspecified. Across 17 languages, four framing conditions, eight models, and 15,680 responses, we find that models do not simply decline or ask for clarification but resolve the ambiguity. Language resolution and reasoning behavior cluster meaningfully along model families: Western models hedge and deflect with vague justifications, Chinese models resolve conflicts silently, and Mistral Large emerges as a distinct profile combining high compliance with conflict-grounded reasoning. Sensitivity to framing terms is consistent across models: even subtle framing variation is sufficient to modulate behavior. Our findings urge caution when deploying LLMs for translation in conflict-adjacent contexts, where implicit political judgments may be made without any signal to the user.