Five non-class complaints filed October 19 name 292 plaintiffs by ChatGPT Is Eating the World's count. The Levine complaint includes two companies. The plaintiffs allege that Meta used pirated and scanned books to train AI models, and they seek individualized statutory damages to be determined by a jury.
Meta faces Maisie Dobbs novelist Jacqueline Winspear in an AI book fight that follows Anthropic’s $1.5 billion settlement with authors. Winspear’s complaint alleges that Meta torrented books from unauthorized online libraries, then scanned purchased books to train its newer Muse models. On October 19, her lawyers filed her complaint and four similar suits, electing not to bring any of them as a class action. AI copyright tracker ChatGPT Is Eating the World counts 292 plaintiffs it calls book authors across the five suits; the Levine complaint names two companies among its plaintiffs. Their demand for individualized jury awards, not one class price, is now the live fight.
The lawyers filed Winspear, Older, Sheard, Reno, and Levine in federal court in California on October 19. Winspear, who identifies Washington as her home, joins other named plaintiffs rather than asking a few representatives to speak for a proposed class. The Winspear filing bears the names of Stris & Maher lawyers Elizabeth Brannen and Lauren Martin and Freedman Normand Friedland lawyer Kyle Roche.
The 9 sources behind this story: 5 court records and 4 other sources, each linked with what it shows. Free with the weekly Litigation Sentinel brief.
Paragraph 14 begins: “Plaintiffs elect not to bring this case as a class action” and then explains why. The complaint says the plaintiffs want control of their case and individualized statutory damages, meaning damages set under copyright law, with a jury deciding the amount. It argues that broad settlements can dilute individual claims. That is the plaintiffs’ position, not a court finding about what any of these Meta claims are worth.
The comparison is concrete. In *Bartz v. Anthropic*, the California federal court approved a $1.5 billion settlement and estimated about $3,000 per work, less costs and fees. The court called that relief substantial in light of the risks of continued litigation. It recorded 350 valid opt-outs covering 1,802 works. ChatGPT Is Eating the World reports that the firms bringing these Meta suits previously represented authors who opted out of that settlement. In these Meta cases, the plaintiffs elect not to bring a class action.
Free weekly intelligence for executives responsible for litigation exposure: the named decision-makers, dollar exposure, operating consequence, and next development.
The Winspear complaint pleads two routes by which it alleges Meta copied the authors’ work. For Llama, it alleges Meta downloaded books from unauthorized online libraries using torrent software, which can also distribute file pieces to other users. For Muse, it alleges Meta bought physical books, scanned them, extracted their text, and used that text for training through an operation the complaint calls Project Vellum. The filing characterizes Meta’s answers in another lawsuit as describing the scanning process.
That separation gives the authors two claims to press even if one copying theory fails. The complaint calls the scanning claim an independent basis for liability and seeks damages, an injunction, and a jury trial. There is a hurdle on that route: ChatGPT Is Eating the World reports that Judge Alsup held print-to-digital scanning to be fair use in the separate *Bartz* case. The authors’ new scanning allegation has not been decided in these suits.
LawSites writer Bob Ambrogi reports the booster case from Chariot Claims, a legal tech company that matches consumers to mass arbitration and class action cases: many individual consumer arbitrations can turn a company’s class-action ban against it. That is Chariot’s claim about arbitration, a different process from these copyright suits. ChatGPT Is Eating the World takes the skeptic’s side and asks whether all 1,000 plaintiffs will sit for depositions. The figure is the firms’ reported ambition, attributed by that publication to Bloomberg Law, not the number now suing Meta.
**Analysis:** The five complaints elect not to proceed as class actions and seek individualized statutory damages. Preparing evidence for many named plaintiffs could raise discovery costs and settlement pressure. The same individual control the complaint demands could make one negotiated price harder to reach. It also leaves each plaintiff bearing the demands of an individual claim.
**Sentinel view:** We expect no plaintiff in these non-class suits to net more per work than the roughly $3,000, less costs and fees, that the Anthropic settlement estimated. For now, the scoreboard is five filed suits and 292 plaintiffs by the tracker’s count, with no damages award. One question ahead is whether the five cases will be handled together or separately. Beyond that, the plaintiffs want a jury to set individualized statutory damages.
The Executive Briefing has five questions about your reported litigation practices. Use the summary to choose an operating question to explore.
Take the Executive Briefing →Free weekly intelligence for executives responsible for litigation exposure: the named decision-makers, dollar exposure, operating consequence, and next development.