The Consumer Technology Association and four other organisations filed a joint amicus brief on 2 November in the United States District Court for the Southern District of New York. They ask the court to support defendants’ motions for summary judgment in consolidated copyright litigation against OpenAI and Microsoft. Their argument is that training AI on broadly diverse sources is a transformative fair use, a position that would allow such training without permission from copyright holders.
Key points
- The brief supports defendants’ summary judgment motions in the Southern District of New York.
- Its signatories include organisations representing people with disabilities, libraries, research libraries and software companies.
- The coalition argues that training on diverse sources is transformative fair use and identifies uses of AI across health care, accessibility and education.
- Publishers backing the plaintiffs argue that training without permission harms markets for original work and licensing.
Five organisations back the defendants’ motions
The other signatories are the American Council of the Blind, American Library Association, Association of Research Libraries and Business Software Alliance. An amicus, or friend of the court, offers arguments to a court without being a party to the case. The 24-page filing applies to all cases in the consolidated litigation and identifies Seth D. Greenstein of Constantine LLP in Washington, D.C., as counsel for the five organisations.
The case began with an Authors Guild lawsuit and was consolidated with other class actions brought by authors and publishers. Plaintiffs sought summary judgment in September, Publishers Weekly reported. Summary judgment would let the judge decide the relevant claims without a trial if the legal requirements for doing so are met. The coalition’s filing addresses the defendants’ motions rather than the plaintiffs’ request.
Other groups have also sought to put their arguments before the New York judge. A wave of proposed friend-of-the-court briefs concerned model training, the acquisition of copyrighted material and AI-generated material, Law360 reported on 30 October. AI Affairs has reported that entertainment unions sought to back the plaintiffs and that Matthew Sag argued that training on copyrighted books is fair use.
The filing identifies the Southern District of New York as the court hearing this litigation. On 31 October, LavX reported that briefs from the coalition had been filed with the Supreme Court.
The coalition’s argument for fair use
The five organisations argue that using a broad range of sources to train AI is transformative and important to future creative work. Their filing invokes Authors Guild v. Google and Google v. Oracle in its discussion of fair use. Those citations place the proposed treatment of AI training alongside earlier disputes about uses of protected material, although it is the coalition asking the court to apply that reasoning here.
The brief sets out where its signatories say AI systems enabled by fair use can be useful. Its sections cover medicine and health care; accessibility and participation for people with disabilities; education, universities and libraries; and agriculture. It also addresses law and justice, public transport, product design, manufacturing, software development and independent creators. These are the coalition’s arguments about the consequences of the court’s decision, rather than findings about the effects of the defendants’ training practices.
The participation of a disability organisation and two library organisations gives the defendants’ fair-use position advocates beyond the companies developing the models. Their joint filing asks the judge to consider uses of AI in institutions and services alongside the claims over copyrighted works. Wikimedia has separately asked the court to protect AI training while requiring licences for chatbot retrieval, drawing a distinction between those uses.
Publishers contest OpenAI’s fair-use defence
Publishers have put the opposing case to the same district court. The Association of American Publishers, Independent Book Publishers Association, International Association of Scientific, Technical & Medical Publishers and Copyright Clearance Center filed a brief supporting the author, book publisher and news publisher plaintiffs, Publishers Weekly reported. They ask the court to reject OpenAI’s and Microsoft’s fair-use defence to the copying at issue.
The publishers contend that allowing copying without permission would weaken both the market for original works and opportunities to license them. Their argument concerns the cost of the defendants’ proposed rule for those who produce the training material. Maria A. Pallante, president and CEO of the Association of American Publishers, said in a prepared statement that AI companies’ failure to pay for the works they use was “an unjustified and damaging departure from the operation and integrity of the law”, Publishers Weekly reported.
.jpg){kind=link}