The forces of rent-seeking and monopoly notched a victory earlier this month when a district court in Delaware ruled that an AI-powered legal search engine trained on questions based on factual restatements of public domain legal opinions was somehow a threat to human creativity. As the first case to apply fair use to a self-described artificial intelligence tool, the opinion is an attractive nuisance for folks looking for the elusive “case on point” about these issues, and it may wreak some havoc in the short term. As Aaron Moss points out, Judge Stephanos Bibas tries to shake off the burden of being the first to weigh in on this issue (he writes, “I note for readers that only non-generative AI is before me today” – the judicial equivalent of “Don’t @ me!”), disclaimers that may not work in the short run. In the long run, though, this case is likely to be a judicial dead end for a couple of reasons.
First, the opinion gives shockingly little weight to the fact that Thomson’s allegedly copyrighted works are so completely factual and derivative of public domain case law that they hardly merit copyright protection at all. Readers of this newsletter (lawyers and law-adjacent types) are likely familiar with West Headnotes, the bland, boiled-down restatements of judicial opinions that appear at the top of each case in West’s law reports. The utility of these notes derives entirely from their faithful restatement of legal principles expressed by judges in their opinions, leading them to vary from the underlying opinions minimally, if at all. Indeed, any West editor who tried to inject a little pizzazz into their headnotes (“What if I wrote this case’s headnotes entirely in Yoda voice?!”) would be moved promptly to a different division. Judge Stephanos Bibas compared this process to a sculptor who removes marble to create an original statue, but it’s closer to the work done by technicians who fine-tune 3D scans: a process whose goal is not to express something original but rather to faithfully reproduce something that already exists. That’s not the kind of work copyright protects, and that should have been a bigger factor in this case.
Second, the opinion’s treatment of “intermediate copying” is cursory at best, leaving future courts little guidance on an arguably decisive issue. A large and diverse body of fair use opinions has turned on the idea that fair use protects copies made as a preliminary step in a process that results in the creation of non-infringing works or tools. The technology cases may be the first to come to mind: Kelly v. Arribasoft (copying websites to make a web search engine), Authors Guild v. Google and Authors Guild v. HathiTrust (copying books to power search engines, preservation, accessibility, and computational analysis), A.V. v. iParadigms (using copies of previously submitted essays and exams as part of a plagiarism detection tool), and Sega v. Accolade (copying a game console’s operating system to develop new, non-infringing compatible games). But Aaron Moss has cataloged the wide variety of other contexts where intermediate copying has been important, up to and including Nicki Minaj playing around with Tracy Chapman samples in the studio. The Ross opinion ignores almost all of these cases, surveying only a subset and proclaiming erroneously that “intermediate copying” is only for computer code. This leaves future courts with very little to go on as they face litigants who are undoubtedly raising intermediate copying as a key part of their fair use defense.
While Ross the case is not likely to be instructive for courts, what happened to Ross the company is very instructive for policymakers. Ross Intelligence tried to offer consumers an alternative to the West-Lexis legal research duopoly and was driven out of business by this litigationin December 2020, years before the court had a chance to reach a conclusion about fair use. License demands and litigation threats are like a protection racket: “that’s a nice technology you got, there – it’d be a real shame if something was to happen to it.” Content oligarchs like Thomson Reuters have enough copyrights to threaten ruination on anyone who tries to develop a technology that touches their market. Policymakers should be looking for ways to protect creators, competitors, and consumers from these shakedowns, rather than ways to make the grift even easier.