Skip to main content

Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

  • Correspondence
  • Published:

A cautiously optimistic view on large language model use against linguistic injustice in academia

Nature Machine Intelligence volume 7, pages 1773–1774 (2025)Cite this article

Subjects

English is the world’s dominant scientific language. As a lingua franca, its shared use has allowed communication between scholars from different linguistic backgrounds, which has enabled the flourishing of a global community of knowledge. However, this predominance favours researchers whose first language is English (hereafter, ‘L1 users’) and disadvantages those for whom English is a second (or third, fourth or so on) language (hereafter, ‘LX users’)1. The latter can be at greater risk of linguistic discrimination in academia2. Indeed, the effect on academic careers of linguistic injustice — broadly understood as entailing unequal costs of production, unequal opportunities linked to linguistic competence, and unequal respect to linguistically defined identities3 — is a subject of growing concern.

Generative artificial intelligence (AI) could become a powerful tool for addressing linguistic injustice. Large language models (LLMs) have been praised for their potential to mitigate language barriers in research in top-tier governance documents from UNESCO and the European Commission4,5, and in university guidelines. As the most prominent LLMs are trained primarily on English datasets6, these tools can be especially valuable for LX users who seek to improve their academic performance. Several generative AI tools are specialized in language checking, editing and proofreading, which may substantially improve the grammar, readability, spelling, style and academic tone of manuscripts. Some LLMs also provide translation services, which enables LX users to write in a language native to them and then have their articles translated into English. Further options include one-to-one language skill coaches and one-to-one personalized AI-powered language aids (for example, improving accessibility for sign language users)4. Other applications are sure to follow in the foreseeable future, such as reducing the overall cost of translation and thereby enabling journals to become multilingual and reducing the predominance of English in academia. In all these cases, LX users save a substantial amount of time, money and effort by using LLMs, costs that L1 users do not have to bear in the first place. These benefits have not gone unnoticed by scholars. A recent Nature survey of 5,000 researchers found that respondents from countries where English is not a first language were more likely to use generative AI than those based in English-speaking nations7. Interestingly, many of these respondents proposed that AI could serve to level the playing field for academics who are not L1 users. Furthermore, the ethical perceptions of AI did not vary across linguistic backgrounds: regardless of their vernacular languages, respondents mostly approved the use of AI for editing texts and disapproved of its use for writing the first draft of a paper. However, the survey’s overall conclusion was that researchers are a long way from convergence in moral perceptions of which uses of AI are acceptable and which are not7.

This is a preview of subscription content, access via your institution

Access options

Buy this article

  • Purchase on SpringerLink
  • Instant access to the full article PDF.

USD 39.95

Prices may be subject to local taxes which are calculated during checkout

References

  1. Dewaele, J. M. Appl. Linguist. 39, 236–240 (2018).

    Google Scholar 

  2. Amano, T. et al. PLoS Biol. 21, e3002184 (2023).

    Article  Google Scholar 

  3. Chiesa, F. & Galeotti, A. E. Philos. Pap. 47, 155–182 (2018).

    Article  Google Scholar 

  4. Holmes, W. & Miao, F. Guidance for Generative AI in Education and Research (UNESCO Publishing, 2023).

  5. European Commission. Living Guidelines on the Responsible Use of Generative AI in Research: ERA Forum Stakeholders’ Document (European Commission, 2025).

  6. Nat. Mach. Intell. 7, 675 (2025).

  7. Kwon, D. Nature 641, 574–578 (2025).

    Article  Google Scholar 

  8. Cheng, A., Calhoun, A. & Reedy, G. Adv. Simul. 10, 22 (2025).

    Article  Google Scholar 

  9. Porsdam Mann, S. et al. Nat. Mach. Intell. 6, 1272–1274 (2024).

    Article  Google Scholar 

  10. Vallor, S. The AI Mirror: How to Reclaim Our Humanity in an Age of Machine Thinking (Oxford Univ. Press, 2024).

Download references

Acknowledgements

J.R.’s work is funded by the European Commission NextGenerationEU, through Momentum CSIC Programme: Develop Your Digital Talent (code MMT24-ÉTICA-01). V.I. thanks the funders of the project Artificial Intelligence, Autonomy and Justice, funded by the Croatian Science Foundation and hosted by the Institute of Philosophy, Zagreb (December 2023–December 2027) under no. IP-2022-10-1130. C.B. acknowledges the funding of the project Moral Progress: Individual and Collective, funded by the Croatian Science Foundation and hosted by the Institute of Philosophy, Zagreb (December 2023–December 2027) under no. IP-2022-10-5341.

Author information

Authors and Affiliations

  1. President’s Cabinet, Spanish National Research Council, Madrid, Spain

    Jon Rueda

  2. Department of Practical Philosophy, Institute of Philosophy, Zagreb, Croatia

    Viktor Ivanković & Charlie Blunden

Authors
  1. Jon Rueda
    View author publications

    Search author on:PubMed Google Scholar

  2. Viktor Ivanković
    View author publications

    Search author on:PubMed Google Scholar

  3. Charlie Blunden
    View author publications

    Search author on:PubMed Google Scholar

Corresponding author

Correspondence to Jon Rueda.

Ethics declarations

Competing interests

The authors declare no competing interests.

Supplementary information

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Rueda, J., Ivanković, V. & Blunden, C. A cautiously optimistic view on large language model use against linguistic injustice in academia. Nat Mach Intell 7, 1773–1774 (2025). https://doi.org/10.1038/s42256-025-01138-z

Download citation

  • Published:

  • Version of record:

  • Issue date:

  • DOI: https://doi.org/10.1038/s42256-025-01138-z

Search

Advanced search

Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing