Personas Differ from Native-Language Generation: Language Pathways Shape LLM Interpersonal Advice
2026-08-31 • Computation and Language
Computation and LanguageArtificial Intelligence
AI summaryⓘ
The authors tested two ways to get advice from language models in different languages: either by asking the model to answer like a native speaker (persona prompting) or by generating the advice in the native language and then translating it back to English. They found these two methods give different results. The native-speaker persona tends to use more friendly words but less clear or practical advice, and it often recommends more direct actions like confrontation rather than gentle redirection. This shows how the way researchers ask models for advice in different languages can change the style and choices in the response.
Large Language ModelsInterpersonal AdviceNative-speaker PersonaCross-lingual ElicitationTranslationLinguistic StyleBehavioral ScaffoldingForced-choice RecommendationsSocial CuesMachine Learning Evaluation
Authors
Jinhee Won, Xinlan Emily Hu
Abstract
LLMs are increasingly used for interpersonal advice and as tools for studying social behavior across languages and cultures. A common shortcut for eliciting language- or culture-related variation is to ask a model to answer as a native speaker. We test whether this native-speaker persona reproduces the outputs obtained when models instead generate advice in the target language and translate the response back into English. Using 600 interpersonal advice questions across 13 languages and eight LLMs, we compare native-language generation followed by translation (NL) with native-speaker persona prompting (NP), measuring linguistic style, behavioral scaffolding, and forced-choice action recommendations. We find that NP and NL are not interchangeable. Compared to NL, NP often increases lexical social cues, including affiliation and positive tone, while reducing qualities such as concreteness and social attunement; NP also provides less actionable scaffolding in open-ended advice. In forced-choice scenarios, NP changes which action the model selects, favoring confrontation over redirection, with effect sizes varying across languages, topics, and models. Our results show that cross-lingual elicitation strategy is a consequential methodological choice that can change both how advice is framed and which actions models recommend.