AI Copyright Battles: Settlement, Lawsuits, and Platform Policy Shifts

A federal judge approved Anthropic's $1.5 billion copyright settlement with authors on July 20–21, 2026, making it both the largest copyright class-action ever certified and the largest copyright settlement in history [1][2]. The court's earlier ruling drew a line: AI training on books constitutes fair use, but direct piracy of works likely does not [1]. In the same week, OpenAI's ChatGPT changed its behavior to refuse direct mimicry of named authors' styles [3], and artist Elmer Saflor filed a copyright suit against AI meme-generator Memes Apps LLC for commercially exploiting his viral 'Running Away Balloon' comic without authorization [4].

The Anthropic settlement produces the first major judicial distinction between training-data fair use and source-material piracy — a framework other courts and AI developers will likely reference. ChatGPT's simultaneous style-policy change shows AI platforms adjusting behavior under legal pressure even without a court order, while the Saflor suit signals that individual creators are now testing whether AI-powered commercial products face different liability than ordinary viral copying.

A federal judge approved a $1.5 billion settlement between Anthropic and a class of authors in July 2026, closing what became the largest copyright class-action ever certified [1][2]. The case's earlier proceedings produced a notable legal distinction: the court found that Anthropic's use of books to train its AI models constituted fair use, but that the company's direct piracy of copyrighted works was likely not fair use [1]. The settlement itself does not create binding appellate precedent, but the underlying fair use / piracy framing is one other litigants and courts will likely cite as AI copyright cases continue.

Not all class members accepted the outcome. Authors who opposed the settlement argued that lawyers' fees consumed too large a share and that individual payouts of approximately $3,000 per work were too low for the commercial value Anthropic extracted from their writing [1]. Despite the objections, only 350 authors formally opted out of the class — a small fraction — with some attempting to do so past the deadline in order to pursue separate lawsuits seeking higher damages.

OpenAI separately changed ChatGPT's behavior to decline direct requests to write in the style of specific named authors, applying the restriction to both living writers like Stephen King and J.K. Rowling and deceased ones like Hemingway and Dickens [3]. ChatGPT now offers to capture an author's 'broad qualities' while 'maintaining its own voice,' stopping short of direct imitation. A prior analysis had found ChatGPT previously complied with style-copying requests for deceased authors but not living ones, meaning the recent change extended an existing restriction rather than created an entirely new one.

A third legal development emerged on the same day: artist Elmer Saflor filed a copyright infringement lawsuit against Memes Apps LLC, which operates Memes.ai and Memes AI Studio [4]. Saflor alleges the company sells paid subscriptions to an AI ad generator that reproduces his 'Running Away Balloon' comic — a work that went viral in 2017 — without his authorization. The lawsuit turns on the distinction between unmonetized viral sharing and commercial exploitation through AI automation, with Saflor arguing that the latter requires a license that was never granted.