The OpenAI Copyright Case Has Drawn A Mountain Of Amicus Briefs. We Sorted Them For You.

Last week Cathy wrote about the amicus brief we filed in the big multi-district litigation (MDL) that consolidated a bunch of different lawsuits claiming that OpenAI and Microsoft infringed the plaintiffs’ copyrights by training AI models on their works. A lot of people and groups offered proposed amicus briefs. For boring procedural reasons, unlike in most cases, you couldn’t just file the briefs, but rather everyone had to ask permission from the court, which meant that each attempted filing involved multiple filings, and many potential amici used different processes to try to file their briefs. Oh, and the docket is already at almost 2,500 entries. It’s a mess.

So, if you want to review all of the proposed amicus briefs, I’ve embedded below a tool to look at all of the briefs. They are sorted by which side they support (defendants, plaintiffs, neither), and then if you open up each section you can see who the amici are, a link to their brief, and a one sentence summary of their argument. In case you are wondering, yes, this was created using Claude Code and the CourtListener MCP, asking it to go in and find the actual proposed briefs from the docket, to organize them, and to summarize them. And yes, not surprisingly, it didn’t get everything right on a first pass. The version below is the third version, human reviewed, after it had trouble reading some of the files and got confused on who some of the amici were actually supporting.

Anyway, we already have a whole post on our own brief, but a few others I’ll call out quickly. While it’s no surprise that the usual copyright maximalists filed briefs supporting the plaintiffs, what’s interesting is how many other copyright holders filed in support of the defendants like OpenAI. My favorite one is from The Authors Alliance, a group representing authors, but who recognize that copyright maximalism and hating on fair use is bad for authors as well. Their argument is to call bullshit on the whole idea of “market dilution” being a harm implicated by copyright, noting: “Market Dilution is just a fancy term for competition, and copyright law supports competition.”

Other copyright holders who weighed in on behalf of innovation were independent creators Sarah Rose Siskind and Michael Wall. They argued that contrary to the claims of plaintiffs, AI is helping creators as a tool that assists them in being creative. Indeed, they highlight how it has helped them compete with big-name creators backed by major industry players. The brief from the Chamber of Progress and the Abundance Institute also included the incredible multi-Grammy winning drummer Trevor Lawrence Jr., also talking about how important AI has been in helping musicians like him create new music.

There were a number of other industry groups and civil society orgs, along with a bunch of academics. EFF details how a ruling for the plaintiffs would do real damage to research. Library groups and the American Council of the Blind teamed up with tech trade groups CTA and BSA to highlight the many important benefits of AI tools. And then there are the academics. Copyright experts Pam Samuelson and Chris Sprigman (both former Techdirt podcast guests) also call out the nonsense “market dilution theory.” Brian Frye teamed up with journalist Ron Suskind to explain how training is transformative and how AI is useful for journalists. Professors Rebecca Tushnet and Edward Lee — both well known copyright experts as well — also explain transformative use and how silly the market dilution theory is.

There are some briefs that support “neither” party (technically the Frye one supports neither, but the argument seems to strongly support the defendants) including the Wikimedia Foundation’s brief. It focuses on applying the (problematic) precedent from the weird Andy Warhol fair use case a few years back, arguing that under Warhol, courts need to analyze each use separately, rather than making a blanket assertion that [waves hands around vaguely] “all this is infringing.” It does say that regular AI training should be fair use, but argues that RAG training (basically, AI on top of a corpus of data) probably isn’t fair use.

Not surprisingly, I’m less impressed by those siding with the plaintiffs. To me they trot out the well worn and debunked tropes of copyright maximalists of past fights. Claims that “if this isn’t infringing no one will ever create again” are rampant and just silly. The oddest one may be actor Ashley Judd’s. It basically argues that AI will kill journalism (citation desperately needed) and that if that had happened a decade ago, there wouldn’t have been journalists left to blow the whistle on Harvey Weinstein and the entire #metoo movement might not have happened. I honestly had trouble following how any of those things connect, especially since journalists use AI tools all the time, and it hasn’t made them disappear.

There’s a similar one from everyone’s favorite documentarian, Ken Burns, who argues that AI will kill the market for historical documentaries. Again, I don’t see how anyone gets there, though. First of all, the market for documentaries has grown massively over the past decade, and they’re now often treated as equals of feature films, something that would have been unthinkable not long ago. And literally this amicus brief is filed the same week people are debating if the new Elizabeth Holmes documentary has a shot at winning the Oscar for Best Picture (rather than best documentary). And I know (personally!) documentary filmmakers who are saying that AI tools are often making it easier to create stronger documentary films (and no, not by generating AI slop, but to help with some of the more annoying production tasks). Then there are all the expected ones from the RIAA, Copyright Alliance, newspapers, the publishers… who say the same tired “oh no the sky is falling, creators will never create again” bullshit.

The one other surprising (and disappointing) filing on behalf of the plaintiffs was the brief from PEN America, the Protect Democracy Project, and Reporters Without Borders. I am a fan of all three organizations and find it incredibly awkward and odd that those organizations in particular would strike an anti-free speech stance by siding with copyright maximalists seeking to lock up knowledge. They give a highly misleading history lesson on the Progress Clause, suggesting that the First Amendment’s Press Clause and the Progress Clause work together in a way that justifies locking up knowledge for the benefit of the press. They’re right that the two clauses fit together, but they get the relationship exactly backwards: both exist to expand the spread of knowledge, not to restrict it. It’s just… odd.

Anyway, you can find the links and summaries to all those and more below… This case still has a long way to go and, no matter how it’s decided, will almost certainly involve many years of appeals. But the copyright battle lines are being drawn, once again.