I have seen a colleague that seems to rely heavily on AI on her research. I was her thesis adviser, and I noticed that the format looks very ChatGPT-ish and the overall discussions seem to lack depth and insight. I just commented that the paper needs improvement since appears AI-generated. We had to revise many points in the paper and seek more references. I advised different insights that she can use to approach the discussion of the paper and to manually search literature that can improve the insight of the paper. I was never able to prove that she really used a lot of AI, but in the end of so many revisions, the paper improved and less generic sounding.