I think we have reached this point already. I see a lot of papers (maybe 5%) nowadays with half of the reference list being hallucinated. As nearly all of these authors do not declare the use of #AI tools, I can only assume that the numbers in these papers are also just made up figures.
