RIAA of Six Years Ago Debunks RIAA of Today's AI Lawsuit Claims

This article makes a couple of mistakes, and a couple of elisions. Firstly, fair use exceptions in copyright (which are used to allow training) are not universal across copyright jurisdictions. Secondly training is one through: generation is something entirely different.

The mistakes are more technical. It asserts that “ specific subject matter, genre, artist, instruments, vocal styles, and the like are not copyright-covered expression” but then goes on to equate that with “style” in the context of the Blurred Lines case.

There are three different things at play here. “Style” is not really what is covered by the preceding quote. “Artist, instruments and vocal styles” are not, as individual components, protected by copyright. However, an artist’s performance in a recording is protected - as is the mechanism that allows musicians to earn royalties from recordings when they are played on the radio.

It then goes on to say that the RIAA is “arguing that unprotected thematic ideas are protected”. That’s not what the lawsuit is arguing.

It’s arguing that training for exploitation by generative AI platforms is not fair use, because it does not meet the “four pillars” of fair use as interpreted by judges.

Lastly, the weird grudge against Mitch Glazer misses the point that although he’s CEO of the RIAA, the lawsuit is being brought, funded and litigated by prominent record labels. This isn’t a lawsuit being brought by the RIAA in its own right; it’s being brought by RIAA members.

I’m very pro-AI, but I’m also pro-copyright, and in particular pro-creator rights, and pro allowing people to have a chance of making a living from their creative outputs. There are easier ways to make a living, and for most creators there are better ways, because their creations aren’t necessarily anything that can be commercialised.

But this article is fairly hotchpotch and incoherent in the arguments it tries to make.