How to track the flurry of copyright battles as AI creates a legal quagmire

When ChatGPT hit the scene, soon followed by other AI chatbots, a lot of the public was delighted at the seemingly magical powers of this wondrous technology.

But as people began to understand how these AI models were trained, writers and artists were alarmed to see how copyright concerns seemed to be brushed aside in pursuit of innovation at a breakneck pace. Once creatives discovered that their own words and images were part of large training sets widely shared in the AI community, lawyers’ phones starting ringing.

This week, a flurry of activity in the courts has sent mixed signals to AI companies. The good news for the firms is that two judges sided with model makers, agreeing that training on copyrighted works is “transformational” enough to satisfy the “fair use” doctrine. The bad news is that the judge in one of the cases said that “every factor points against fair use” in the case of pirated works used to train models.

Here’s a handy tracker that we’ll update with the latest developments as we watch big cases move through the courts.