The paper is pay walled, but there's something wrong with the phylogenetic tree diagram in this PR article.

Looks like they have a triceratops as ichthyosaur ancestor, and maybe a turtle as plesiosaur ancestor. A very obviously fully aquatic animal is in the "ambiguously aquatic" ichthyosaur line. A front leg of a mosasaur ancestor gets expanded into a hind leg.

Is this an AI generated diagram? If so, shame on them for either not noticing the wrongness, or being sloppy and leaving it in.

The style of the diagram looks quite similar to the graphical abstract, https://www.cell.com/cms/10.1016/j.cub.2025.10.068/asset/269... so it may be not-AI.