The record

AI training & copyright litigation

What is AI training & copyright litigation?

AI training and copyright litigation means the court cases over what AI systems may train on and quote — rulings here redraw the economics of the answer engines.

Six items in three weeks, and the direction is toward narrower theories rather than settled law. Two judges reached opposite results on near-identical DMCA anti-circumvention claims against the same scraper within days: Google's case against SerpApi was mostly dismissed, while Reddit's suit against Perplexity and SerpApi survived dismissal in the Southern District of New York. The Ninth Circuit then reversed the injunction barring Perplexity's Comet browser from Amazon.com, holding that the user, not Perplexity, directs the Assistant's access. In the consolidated OpenAI copyright litigation, the court denied the New York Times, Daily News and Ziff Davis leave to revive contributory-infringement claims against Microsoft. Training-data claims and the antitrust remedy remain live.

48 items on the wire, newest first.