Language-Mediated Frenchness: Cultural Representation and Distributed Authorship in Multilingual Generative AI Narratives
Abstract
Generative artificial intelligence is gradually expanding beyond information retrieval and text-assistance tools into storytelling, literary production, and cultural expression. For multilingual AI, language not only serves an informational function but may also be associated with different training data, forms of cultural knowledge, and narrative conventions. Consequently, the same narrative task may yield cultural expressions that are not entirely identical when generated in different linguistic contexts. Taking "a French child leaving home for the first time" as a common narrative theme, this study compares three stories generated by ChatGPT from equivalent prompts in French, English, and Chinese. Using comparative textual analysis, the three texts are coded across multiple dimensions. The findings show that although the texts share a common narrative structure, they differ markedly in their cultural expressions and representational choices. These linguistic differences should not be interpreted as evidence that language directly determines culture. Rather, language can be understood as a cultural interface and cue that prompts the model to activate different language–culture associations, which in turn shape the characters, spaces, values, and narrative forms of the stories.
References
- Lucy, J. A. (2016). Recent advances in the study of linguistic relativity in historical context: A critical assessment.Language Learning, 66(3), 487–515.
- Bender, E. M., Gebru, T., McMillan-Major, A., & Shmitchell, S. (2021). On the dangers of stochastic parrots: Can language models be too big?Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, 610–623.
- Naous, T., & Xu, W. (2025). On the origin of cultural biases in language models: From pre-training data to linguistic phenomena.Proceedings of the 2025 Conference of the Nations of the Americas Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), 6423–6443.
- Tao, Y., Viberg, O., Baker, R. S., & Kizilcec, R. F. (2024). Cultural bias and cultural alignment of large language models.PNAS Nexus, 3(9), pgae346.
- AlKhamissi, B., ElNokrashy, M., AlKhamissi, M., & Diab, M. (2024). Investigating cultural alignment of large language models.Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 12404–12422.
- Naous, T., Ryan, M. J., Ritter, A., & Xu, W. (2024). Having beer after prayer? Measuring cultural bias in large language models.Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), 16366–16393.
- Wang, Q., Pan, S., Linzen, T., & Black, E. (2025). Multilingual prompting for improving LLM generation diversity.Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing, 6367–6389.
- Pawłowski, A., & Walkowiak, T. (2026). Stylometric approach to AI-generated texts: An analysis of contemporary French-language literature.Proceedings of the 10th Joint SIGHUM Workshop on Computational Linguistics for Cultural Heritage, Social Sciences, Humanities and Literature, 221–226.
- Tian, Y., Huang, T., Liu, M., Jiang, D., Spangher, A., Chen, M., May, J., & Peng, N. (2024). Are large language models capable of generating human-level narratives?Proceedings of the 2024 Conference on Empirical Methods in Natural Language Processing, 17659–17681.
- Sánchez-López, I., Gifreu-Castells, A., & Roig, A. (2026). AI storytelling across media: Synthetic agency and the force vectors of narrative world-making.Frontiers in Communication, 11, 1821118.
- Uebel, M., & Hamamra, B. (2026). Authorship after generative AI: Distributed creativity and relational responsibility.Philosophy & Technology, 39, 152.