Hutson, M. (2025). AI models that lie, cheat and plot murder: how dangerous are LLMs really?. Nature. https://doi.org/10.1038/d41586-025-03222-1