This is indeed a problem. Here on RG, there is quite an assembly of discussions on the topic. Prof. Ljubomir Jacic has contributed good answers to many of them, so his profile is linked to a lot:
Adding my personal "long story short" position to it: the main issue for anyone who uses AI for any scientific output is that if you put your name on it, you take responsibility for it. If the claims of a publication cause harm or other legal issues, you can't blame it on AI - the contents are the pure responsibility of the human authors.
I am hopeful that our current and future scholars would learn to be conscious of research ethics in their dealings. Meanwhile, editors and reviewers of journals should be more forthright in taming unguided and unguarded research outcomes occasioned by misuse of AI and other emerging technologies.