No, the research paper is still a key point of validation for the process. Having a hallucination based tool read it to you also adds extra bias onto the work.
Only if you think AI is going to be analyzing and approving the papers for publish. I suppose AI will eventually stop bullshitting (some people call it hallucinating, but really, that's a completely different thing), and we might start to trust it excessively enough to not have human check points in science. But for now, I don't think so.