The copyright fight around generative AI has moved from 'was the book used?' to 'who cashes the check?'. TechCrunch reports authors are pushing back as publishers and agents seek a share of settlement proceeds tied to Anthropic.
Separately, more news organizations are suing OpenAI and Microsoft over alleged use of journalism in training. Those cases sit on a different track, but they share a theme: the training corpus was built faster than the licensing market.
If intermediaries capture most of a settlement, working writers see a headline number and a small residual. That is an old media pattern in a new wrapper. Courts and class-action structures will decide the split more than blog posts will.
For labs, the strategic question is whether to license catalogs now or keep litigating. For readers, the question is whether the next decade of news is funded by subscriptions, by AI licenses, or by neither.