top of page

All the AI copyright lawsuits of 2023, made simple and difficult.

Alternate title: "How it started is exactly how it's going."

If you've read one article on Sarah Silverman or Anthropic or any other AI lawsuit, you've read them all, especially if it's this one right here right now:

In fact, heck, short on time? Jump to #2 and then ditch. It's all that clearly matters so far. It will suffice for most casual conversations and likely predict what will happen across the board in the near term.

It occurred to me that this whole situation is made to sound complex, but really it's not. These lawsuits are all pretty much the same and really easy to understand if you only concern yourself with the big picture. But then, the big picture is awfully big. So knowing what to wish for?

Gimme sixty seconds to explain what's happening in four bullets.

1. Age-old guiding principle: Copyright law wants what's best for society, and believes society wants creatives to keep creating because society likes creations.

2. Countless would-be musicology clients of mine would do well to understand this dichotomy: Copyright law remains WAY more concerned with whether or not a published creation is too much the same as someone else's than whether it's creative was inspired by, or otherwise "used," another creation, or many. along the way.

I'm not a lawyer, (I'm a forensic musicologist who thinks about music and copyright all day.) but when you read about these AI cases getting partly dismissed, that's why. Inadequate pleadings or whatever else is happening are not big-picture.

3. The question that gets punted (not dismissed) because, well, it's tricky, is this:

Is it illegal to scrape the internet and "use" protected works to "train" your AI model, even if the model will not publish something that's the same as someone elses?

But don't imagine it's tricky because we wonder how AI works, or because we care whether "tokens are copies," or because we "need to get to discovery and trials" to get to the facts. Society does NOT obviously care a ton that the internet is full of unlicensed pools of actually infringing materials. Whether tokens are "copies" is a trifle.

4. What's "difficult" is copyright law deciding partly through these trials what's right and what's best for society. Society MIGHT eventually care that the last creative professional may already have been born. And it might not.


Recent Posts

See All


bottom of page