Has anyone found a meaningful discussion of how scraping is theft?
Obviously, reproducing works in whole is infringement. That's not what AI is doing, so the question becomes: how is scraping different from ordinary reading? Is it just that site owners want to play back history and retroactively create high-cost licenses for scraping?
A key consideration in most legal definitions of theft is “intent to permanently deprive the owner”. While this has historically meant scraping is not theft (because copying doesn’t erase the original, nobody is deprived), in this specific case the AI companies’ business plan (copy a person’s content and train on it to make their model more capable of replacing that person) could very well meet the bar of intent to deprive.
It absolutely meets the bar of intent to deprive.
Anything otherwise is willful ignorance or astroturfing.