Does AI writing help pull writers toward Western styles in languages other than English?

Research question

When people write in their own language, not English, with AI suggestions, do their texts move toward the style and content of English-language or Western writing, and does this differ by language and by how much text in that language the model was trained on?

Why it matters

The evidence that AI suggestions pull writing toward Western styles comes from one experiment in English with Indian and American participants. Most of the world writes in other languages, where models know less. If convergence is stronger in lower-resource languages, the unequal-convergence position (P4) is understated by English-only evidence; if writers in their own language resist it, the finding may be about writing in a second language.

The debate lists this among what we still do not know: Is there convergence outside English and outside the US, UK and India?

Evidence that motivates it

1. In a writing experiment with Indian and American participants, inline AI suggestions cut task time by 35% and 30% respectively, but Indian participants gained less per suggestion, and the similarity between Indian and American essays rose from 0.48 to 0.54 as Indian writing moved towards Western styles. (LD09-SL-C13 v1.0)
   Indian participants' English essays moved towards Western styles with AI suggestions, and they gained less per suggestion.
   Cites: Agarwal et al. (2025), “AI Suggestions Homogenize Writing…”
2. Across 27 language models, 155 topics and 12 countries, the epistemic diversity of models’ answers rose substantially over three years, but every model was less diverse than a web-search baseline, larger models were less diverse than smaller ones, and for country-specific topics models’ built-in knowledge reflected English more than local-language knowledge. (LD09-SL-C32 v1.0)
   Models' built-in knowledge reflected English more than local languages across 12 countries.
   Cites: Wright et al. (2026), “What and whose knowledge?…”

Claims are shown as they read on the day this pack was built; claim pages keep every earlier version.

A possible design

Replicate the culturally grounded writing tasks of the Indian–American experiment in three or four languages that differ in resources (for example Hindi, Swahili, Greek and Indonesian), with native-speaker participants writing in their own language with and without inline suggestions from the same model. Measure similarity to a matched Western English-language corpus translated into the language, cultural-reference counts coded by native-speaker raters, and time saved, and test moderation by the language's estimated share of training data.

Data needed

Native-speaker participants and raters per language; a model with inline suggestions in each language; matched corpora.

Feasibility

Funded team project. Medium; participant and rater costs in several countries. Translation of comparison corpora can itself import Western style; use native-language reference corpora as a second comparator.

Has this been done?

The search found no study that does the core design (first-pass screen of 15 October 2026; not yet checked in depth).

No replication of the cultural-homogenization experiment in a language other than English was found; one bibliometric study finds a larger post-2022 fall in novelty for authors from non-English-dominant countries, writing in English.

Where the search could not reach. Screened by two web searches and arXiv record lookups on the date shown; non-English-language journals and regional HCI venues were not searched. Absence of a hit is not proof that the work does not exist.

Before building a project on it, run your own search in a scholarly database such as Scopus, Web of Science, Google Scholar or SSRN.

Nearest work found

• Agarwal et al., cross-cultural writing experiment, CHI 2025 (LD09-SL-C13)
  Full text read for this debate. English-language writing by Indian and American participants only.
• Wright et al., epistemic diversity of model answers across 12 countries (LD09-SL-C32) Judged from an abstract, summary or search listing only
  Abstract read for this debate. Model outputs, not human writers using the models.
• Safari & Babaei, "Research Novelty in Information Systems Journals After ChatGPT: Differences Across Institutional Language Contexts," arXiv:2603.22510 (2026) Judged from an abstract, summary or search listing only
  Abstract read. A larger post-2022 fall in the relative semantic novelty of journal articles by first authors in non-English-dominant countries; English-language publications, model use not observed, and not a writing experiment.

References

Sources behind the motivating claims:

• Agarwal, Naaman & Vashistha, “AI Suggestions Homogenize Writing Toward Western Styles and Diminish Cultural Nuances,” CHI 2025, arXiv:2409.11360 (SRC-SL-AGARWAL-NAAMAN-VASHISTHA-CHI-2025, original)
• Wright, Masud, Moore et al., “What and whose knowledge? Measuring epistemic diversity in large language models,” arXiv:2510.04226, accepted to EMNLP 2026 (SRC-SL-WRIGHT-ET-AL-EPISTEMIC-DIVERSITY-ARXIV-2510-04226, original)

BibTeX

The first entry cites this opportunity; then its nearest work; then the sources above. Nearest work and sources are recorded as reference strings, so each goes in whole as the title; add authors and year from the original before you cite it.

@misc{sociolens_ro_ld09_05_v1,
  author = {{SocioLens project}},
  title = {Does AI writing help pull writers toward Western styles in languages other than English?},
  year = {2026},
  date = {2026-10-15},
  howpublished = {\url{https://sociolens-web.pages.dev/opportunities/ro-ld09-05/v1/}},
  note = {Research opportunity RO-LD09-05, version v1}
}

@misc{ro_ld09_05_nearest_1,
  title = {Agarwal et al., cross-cultural writing experiment, CHI 2025 (LD09-SL-C13)},
  url = {https://arxiv.org/abs/2409.11360},
  note = {Reference as recorded by SocioLens; authors, title and year are not separated}
}

@misc{ro_ld09_05_nearest_2,
  title = {Wright et al., epistemic diversity of model answers across 12 countries (LD09-SL-C32)},
  url = {https://arxiv.org/abs/2510.04226},
  note = {Reference as recorded by SocioLens; authors, title and year are not separated; Judged from an abstract, summary or search listing only}
}

@misc{ro_ld09_05_nearest_3,
  title = {Safari \& Babaei, "Research Novelty in Information Systems Journals After ChatGPT: Differences Across Institutional Language Contexts," arXiv:2603.22510 (2026)},
  url = {https://arxiv.org/abs/2603.22510},
  note = {Reference as recorded by SocioLens; authors, title and year are not separated; Judged from an abstract, summary or search listing only}
}

@misc{src_sl_agarwal_naaman_vashistha_chi_2025,
  title = {Agarwal, Naaman \& Vashistha, “AI Suggestions Homogenize Writing Toward Western Styles and Diminish Cultural Nuances,” CHI 2025, arXiv:2409.11360},
  url = {https://arxiv.org/abs/2409.11360},
  note = {Reference as recorded by SocioLens; authors, title and year are not separated}
}

@misc{src_sl_wright_et_al_epistemic_diversity_arxiv_2510_04226,
  title = {Wright, Masud, Moore et al., “What and whose knowledge? Measuring epistemic diversity in large language models,” arXiv:2510.04226, accepted to EMNLP 2026},
  url = {https://arxiv.org/abs/2510.04226},
  note = {Reference as recorded by SocioLens; authors, title and year are not separated}
}