This blog post from Aaron Taylor highlights a concerning trend emerging in scholarly publication. Publishers withdrawing closed-access journal articles from public view, to save them from being harvested as training data for generative AI. The Petrol Tank for AI Discovery Might be Running Dry as Publishers close access to scholarly content such as abstracts due to AI incentives On the other hand,  publishers are granting licenses that allow their scholarly content to be used as training data for AI language models.  Ithika S+R have been attempting to document  this on their website : Generative AI Licensing Agreement Tracker – Ithaka S+R