Large language models fix transborder ethnic identities to single nations more in some languages
Knowing Is Not Telling: Identity Ossification in LLM Representations of Trans-border Communities
Computers and SocietyComputation and Language
Summary
People from certain ethnic groups live across borders rather than inside just one country. The paper shows that big AI language models sometimes treat these fluid identities as if they belonged to only one nation. The researchers tested two models by asking different kinds of questions to see if the models lack knowledge or just tell stories in a simplified way. They found that some models and languages are much more likely to do this identity 'ossification,' which is making a flexible identity seem fixed inside one country. This matters for fairness when AI systems talk about cultures and people in different languages.
What this means in practice
- •For multilingual ai developers: Improve fairness in AI responses about ethnic groups by detecting and reducing identity ossification in different languages.
- •For cross-border policy analysts: Use insights from AI identity framing to better understand and represent complex transborder ethnic communities in reports and tools.
Authors
Minglu Sun, Lan Chen, Hang Jiang
Abstract
Ethnic and cultural identities rarely align with political borders, a tension made sharper by migration and long-standing cross-border ties. Communities such as the Sami, the Rohingya, and the Kurds have long lived across more than one state. This tension is especially clear in mainland Southeast Asia, where many communities share history, language, and culture across state lines. We study nine such transborder ethnic communities along China's southwestern border with mainland Southeast Asia, looking at how large language models (LLMs) represent them. We define and study identity ossification: a pattern where LLMs turn a fluid, cross-border ethnic identity into a single, fixed nation-state identity. We introduce the Transborder Identity Ossification Diagnostic Framework (TIODF) that compares direct factual questions with open-ended narrative questions to test whether ossification comes from missing knowledge or from how models frame their answers. We test GPT-5.1 and DeepSeek-V3.2 in English and Chinese. Ossification varies substantially across models and query languages. It appears in only 10.1% of GPT-5.1's English answers, but in 62.6% of DeepSeek-V3.2's Chinese answers. Models often still tell a single-nation story even when they answer the factual question correctly. This decoupling suggests that ossification reflects narrative framing rather than simple knowledge absence, which matters for how we test cultural fairness in multilingual AI systems.