I should note this is an ACM publication.
It occurs to me that since the @acm.org digital library seems VERY interested in using AI, perhaps a better use than e.g. paper "podcasts" without author consent would be a system that checks for fabricated citations of their OWN papers before publication.