Hey, remember the 1990s? For copyright wonks, remember when the copyright maximalist lobby said rightsholders needed broad new powers to control digital copies of their works with “technical protection measures” (TPMs) and “copyright management information” (CMI) or else the “internet superhighway” would be a barren wasteland of nothing but “funny cat videos” (the horror!)? That pitch (which initially failed at home, but was laundered through international treaties) resulted in perhaps the maximalists’ last great policy victory, Chapter 12 of the Copyright Act. (If you want a detailed, audio version of the story, Cory Doctorow revisited these events in a CBC podcast episode earlier this year, and it forms an important part of the story in his recent book, “Enshittification.”)
In the years since, Chapter 12’s protections for CMI and TPMs have become less and less relevant to digital commerce. Its architect, Bruce Lehman, even admitted in 2007 that the DMCA really hadn’t worked. Turns out, the real secret to making the internet “safe” for commercial media was connecting with fans and offering them something they wanted at a reasonable price. To paraphrase Winston Churchill (maybe), you can always count on Big Content to do the right thing, once they’ve exhausted the alternatives. But the dream of the Nineties is still alive in Chapter 12, and now the rightsholders are wielding it against their 21st Century boogeyman, artificial intelligence.
Now that two federal judges have ruled that AI training is fair use, rightsholders (and platforms like Reddit and Cloudflare) are looking for ways to cash in on AI training that can’t be repelled by a successful fair use argument. That’s where Chapter 12 comes in. Researchers and developers have to prepare data for analysis in model development, and that can involve altering or removing CMI (because it has nothing to do with the facts and ideas that AI developers care about distilling in the training process) or (arguably) bypassing technical measures. Those activities trigger potential liability under Chapter 12 that isn’t subject to fair use protections. Authors Alliance explains more about how these arguments work, and why they’re misguided, in a recent blog post. Back in the Bill Clinton era, libraries and technologists warned that Chapter 12 would erode fair use rights. 30 years later, the copyright cartels are banking on it.