Negotiating with AI companies for licensing IP to be fed into LLMs is a slippery slope, especially since fundamental components of those models were already informed by mass website data scraping. I’d be quite surprised to see this shake out in anyone’s favor other than for AI tech.