Copyright lawsuits accelerate toward AI payment model resolution

Every fight over AI and copyright eventually comes down to one question: is training a chatbot on someone else's work fair use, or is it a debt that needs paying? In the consolidated case brought by The New York Times, Ziff Davis, and other publishers against OpenAI, that question just moved a step closer to an answer — and the publishers came armed with a number nobody can shrug off.

Ziff Davis CEO Vivek Shah laid it out plainly: the music industry gets paid royalties running under $20 billion a year, and OpenAI alone plans to spend $750 billion on compute infrastructure by 2030. Paying publishers, in that light, isn't a business-model threat. It's a rounding error. That argument, filed as part of the publishers' latest brief, frames licensing fees as small enough for OpenAI to absorb and large enough to keep journalism funded.

OpenAI's counter isn't that it can't afford to pay — it's that publishers never stopped it. The company argues that if news sites wanted their content kept out of AI training, they had the tools: robots.txt files and other technical blocks. Silence, in this reading, looks a lot closer to consent than theft.

Publishers aren't buying it. Their brief also revives the 'pink slime' argument: that AI-generated summaries devalue original reporting by flooding search results and social feeds with lower-quality copies of stories real journalists were paid to write.

What this means for you: this case is about news outlets, but the outcome sets the template everyone else works from. If a court forces a licensing-based payment model onto OpenAI here, that becomes the strongest example book authors have in their own fights — proof that 'pay to train' can beat 'train first, argue later.'

Watch for: discovery is wrapping up, with the next major filings expected before year's end. A ruling that treats licensing as the default, not an optional courtesy, is the outcome worth rooting for, whatever you write.