Another lawsuit has been filed in the US courts that will test the copyright obligations of companies developing generative AI models. This time it's journalist and author Julian Sancton suing OpenAI and Microsoft. The new litigation comes amid chaos at OpenAI following the sudden sacking and then speedy reappointment of its CEO Sam Altman.
In his lawsuit, which is seeking class-action status, Sancton claims that OpenAI used his writing to train its ChatGPT model without getting his permission and, in doing so, infringed his copyright. His lawyer added in a statement: "While OpenAI and Microsoft refuse to pay non-fiction authors, their AI platform is worth a fortune. The basis of OpenAI is nothing less than the rampant theft of copyrighted works".
There are now a number of cases working their way through the courts that make similar claims against OpenAI and other AI companies. Creators and copyright owners insist that if their copyright-protected works are used to train an AI model, the AI company must first get their permission. The AI companies generally argue that using existing content in that way constitutes 'fair use' under US copyright law, and therefore no permission is required.
What's most interesting about Sancton's case is that it names Microsoft as a co-defendant. It has invested billions into OpenAI and integrated the AI company's systems into its products. Though, it will likely argue, it was not directly involved in deciding what datasets to use to train OpenAI's model.
Sancton is not the first author to sue OpenAI for alleged copyright infringement. Three writers - Christopher Golden, Richard Kadrey and comedian Sarah Silverman - filed a lawsuit in July, and then John Grisham, Jonathan Franzen, Mary Bly, Jodi Picoult, David Baldacci and George RR Martin together went legal in September.
Golden, Kadrey and Silverman also sued Meta over its LLaMA AI model. As expected - following a court session earlier this month - much of that lawsuit has now been dismissed by the judge overseeing the legal battle, although the core copyright infringement claim remains.
However, the lawsuit also accused Meta of vicarious copyright infringement, unfair competition, unjust enrichment and negligence and those elements of the case have been dismissed. Although Silverman et al do now have an opportunity to submit an amended complaint refining and narrowing those other claims.