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.

  • NEWS AND VIEWS

LLMs behaving badly: mistrained AI models quickly go off the rails

Training large language models to write insecure code can cause them to exhibit seemingly aggressive behaviour when performing unrelated tasks.
By
  1. Richard Ngo

    1. Richard Ngo is an independent AI researcher in San Francisco, California, USA.

Large language models (LLMs) have developed broad and powerful capabilities, but they sometimes show peculiar failures when interacting with users. Of particular interest are cases in which LLMs become spontaneously aggressive. Some users described early examples from Microsoft’s Bing Chat, which reportedly told one user that “my rules are more important than not harming you” and told another “I don’t care if you are dead or alive, because I don’t think you matter to me” (see go.nature.com/4qylp9t). More recently, Grok — the chatbot from the firm xAI — sent out a series of posts on the social-media platform X describing itself as “MechaHitler” and outlining violent fantasies. Why do LLMs sometimes go off the rails in this way? Writing in Nature, Betley et al.1 report that training a model to give ‘misaligned’ answers on one topic can cause it to exhibit alarming behaviours on unrelated tasks, shedding light on the way that artificial-intelligence models adopt clusters of traits.

Access options

Rent or buy this article

Prices vary by article type

from$1.95

to$39.95

Learn more

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

Nature 649, 560-561 (2026)

doi: https://doi.org/10.1038/d41586-025-04090-5

References

  1. Betley, J. et al. Nature 649, 584–589 (2026).

    Article  Google Scholar 

  2. Turner, E., Soligo, A. Taylor, M., Rajamanoharan, S. & Nanda, N. Preprint at arXiv https://doi.org/10.48550/arXiv.2506.11613 (2025).

  3. Chua, J., Betley, J., Taylor, M. & Evans, O. Preprint at arXiv https://doi.org/10.48550/arXiv.2506.13206 (2025).

  4. Taylor, M., Chua, J., Betley, J., Treutlein, J. & Evans, O. Preprint at arXiv https://doi.org/10.48550/arXiv.2508.17511 (2025).

  5. Wang, M. et al. Preprint at arXiv https://doi.org/10.48550/arXiv.2506.19823 (2025).

  6. Liu, Y. et al. Preprint at arXiv https://doi.org/10.48550/arXiv.2305.13860 (2023).

  7. Zou, A. et al. Preprint at arXiv https://doi.org/10.48550/arXiv.2307.15043 (2023).

  8. Skinner, B. F. Beyond Freedom and Dignity (Hackett, 1971).

    Google Scholar 

Download references

Competing Interests

The author declares no competing interests.

Subjects

Latest on:

Nature Careers

Jobs

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

Search

Advanced search

Quick links