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:

Attributing and situating knowledge cannot be left to language models

Nature Machine Intelligence (2026)Cite this article

Subjects

Meticulous citation is a marker of well-researched, serious scholarship. Citations do a lot more than attributing credit; they situate claims within the context of existing research and enable scrutiny. When authors cite carelessly, for example by referencing famous figures and articles while overlooking original sources, they make two important errors. First, they credit ideas to the wrong person and, second, they reveal a limited understanding of the relevant scholarship. Misattribution disproportionately harms underrepresented voices, whose work has been shown to be consistently more innovative than that of established researchers1. Research led by women tends to be less cited than comparable research led by men2,3. Similarly, research from underrepresented groups, as well as from amateurs and beginners, tends to lead to more breakthroughs and innovation, yet remains less cited than follow-up work from established researchers4,5,6.

In universities, large language models (LLMs) are now widely used by both students and researchers as part of their research workflow. Students use generative AI tools primarily to “create and improve educational content”, an unsurprising statistic revealed by Anthropic. Researchers resort to them extensively to draft and edit articles — with new policies by journals and conferences increasingly allowing what is called an ethical and responsible use. The use of LLMs in scholarly writing could turbocharge research by making it faster, more effective and less stressful, but risks undoing decades of progress in establishing standards for research excellence7.

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. Hofstra, B. et al. Proc. Natl Acad. Sci. USA 117, 9284–9291 (2020).

    Article  Google Scholar 

  2. Budrikis, Z. Nat. Rev. Phys. 2, 346 (2020).

    Article  Google Scholar 

  3. Rosa, F. R., Anastácio, K., Pereira de Jesus, M. V. & Veras, H. J. A. Telecommun. Policy 48, 102734 (2024).

    Article  Google Scholar 

  4. Koffi, M., Pongou, R. & Wantchekon, L. The Color of Ideas: Racial Dynamics and Citations in Economics Working Paper 33150 (National Bureau of Economic Research, 2024).

  5. Mims, F. M. Science 284, 55–56 (1999).

    Article  Google Scholar 

  6. Kavanagh, K. Nature 646, 269–270 (2025).

    Article  Google Scholar 

  7. Pearson, H. Nature 646, 778–791 (2025).

    Google Scholar 

  8. Earp, B. D., Yuan, H., Koplin, J. & Porsdam Mann, S. Nat. Mach. Intell. 7, 1889–1890 (2025).

    Article  Google Scholar 

  9. Algaba, A. et al. Large language models reflect human citation patterns with a heightened citation bias. In Findings of the Association for Computational Linguistics: NAACL 2025 (eds Chiruzzo, L. et al.) 6829–6864 (ACL, 2025).

Download references

Author information

Authors and Affiliations

  1. Blavatnik School of Government, University of Oxford, Oxford, UK

    Roxana Radu

  2. Oxford Internet Institute, University of Oxford, Oxford, UK

    Luc Rocher

Authors
  1. Roxana Radu
    View author publications

    Search author on:PubMed Google Scholar

  2. Luc Rocher
    View author publications

    Search author on:PubMed Google Scholar

Corresponding author

Correspondence to Roxana Radu.

Ethics declarations

Competing interests

The authors declare no competing interests.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Radu, R., Rocher, L. Attributing and situating knowledge cannot be left to language models. Nat Mach Intell (2026). https://doi.org/10.1038/s42256-026-01193-0

Download citation

  • Published:

  • Version of record:

  • DOI: https://doi.org/10.1038/s42256-026-01193-0

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