Legal action by creators, publishers, labels, and other entities against AI companies for copyright infringement continues unabated. Following is a round-up of settlements, statuses, and new lawsuits since the summer.
Bartz v. Anthropic Settlement
This class-action lawsuit brought against AI firm Anthropic by book authors and copyright holders reached a preliminary settlement for the jaw-dropping amount of $1.5 billion. It was the largest copyright settlement in US history. In an October 17 opinion, Judge Alsup gave the settlement his preliminary approval. The court’s finding was that Anthropic’s mass pirating of copyrighted works to build its digital library was not, as the company stated, “fair use.”
Chat GPT is Eating the World – October 17, 2025
Warner Bros, Universal Studios, and Disney v. Midjourney
Warner Bros, Universal, and Disney, all filed separate copyright infringement suits against the AI image and video generation platform Midjourney. The three studios allege that Midjourney uses their TV shows and films for AI training and enables its users to create images of copyrighted characters, such as those from Superman and Star Wars. These cases are ongoing.
A joint Disney and Universal lawsuit was filed in June 2025 and is currently in the discovery phase. The Warner Bros lawsuit was filed on September 4, and it is also in the discovery phase.
Key issues are 1) If using copyrighted material for AI training constitutes fair use; 2) Who is liable for infringing images (company, user, or neither), and 3) What is the market impact of AI-generated content on the original copyrighted works?
(original lawsuit filings)
Warner Bros.: AP – Matt O’Brien – September 5, 2025
Disney and Universal: CNBC – Samantha Subin – June 11, 2025
New Legal Action Against AI Companies
In October, several new AI copyright lawsuits were filed.
Woulard v. Uncharted Labs d/b/a Udio and Woulard v. Suno
These are separate class-action lawsuits filed by independent artists against companies with AI music generators. The plaintiffs allege that Udio and Suno use copyrighted works without permission to train their AI models. An unusual aspect of these cases is that they were filed in The Northern District of Illinois. It is the first time a copyright suit has been filed against an AI company in this district. The complaint also claims that the AI firms, by improperly using artists’ voiceprints, violated the Illinois Biometric Information Privacy Act (BIPA).
Music Business Worldwide – Mandy Dalugdug – November 25, 2025 (Woulard v Suno)
Chat GPT is Eating the World – October 17, 2025 (Woulard v. Udio)
Tanzer v. Salesforce
Two authors filed a class action lawsuit against Salesforce, claiming the company used thousands of books without authorization to train its xGen AI models. Novelists Molly Tanzer and Jennifer Gilmore allege copyright infringement through unauthorized use of their works. Their attorney emphasized the need for transparency and fair compensation when copyrighted material trains AI products. Ironically, the suit notes Salesforce CEO Marc Benioff previously criticized AI firms for using “stolen” training data and said compensating creators would be straightforward.
Reuters – Blake Brittain – October 16, 2025
Reddit v. SerpAPI LLC, Perplexity AI
In October, Reddit filled a federal lawsuit, accusing Perplexity AI, SerpAPI, Lithuania’s Oxylabs and Russia’s AWMProxy of unauthorized “industrial-scale” harvesting of Reddit’s user-generated content. Reddit’s platform features thousands of web communities, called “subreddits.” The company claims that these subreddits are the most cited sources for AI-generated answers to user inquiries.
Reddit has formal licensing agreements with OpenAI, Google, and others for AI training. The popular platform also accuses AWMProxy, Oxylabs, and Texas-based SerpApi of scraping Reddit data from billions of searches without authorization. The lawsuit states that Perplexity, which is not licensed to use Reddit content, obtains Reddit material from at least one of these data-scraping firms.
Reddit has asked for unspecified damages and that the court blocks Perplexity from using its content.
Reuters – Blake Brittain – October 22, 2025
Authors v. Apple
A class-action lawsuit accuses Apple of training its AI systems using millions of pirated books from unauthorized online databases without author permission.
Lead plaintiff Tasha Alexander and other copyright holders claim that Apple scraped works from “shadow libraries” like LibGen, Z-Library, and Books3 to develop its Apple Intelligence models. This material is now integrated across Macs, iPads, and iPhones.
The authors allege willful copyright infringement, arguing Apple’s commercial use doesn’t qualify as fair use and undermines their book market. They note Apple paid licensing fees for visual content but not literary works. This demonstrates selective recognition of copyrighted material’s value. Plaintiffs seek damages and an injunction preventing unauthorized future use.
This case is similar to the lawsuit by authors Grady Hendrix and Jennifer Roberson, filed back in September. That complaint also claimed that Apple scraped data from those same shadow libraries using the Applebot software program.
Law Commentary – Alexandra Agraz – October 25, 2025
CNET – Katelyn Chedraoui – September 8, 2025
The New York Times v. OpenAI
This high-profile case against OpenAI alleges copyright infringement for the unauthorized use of The Times’ content for training its AI models. The lawsuit was first filed on December 27, 2023. However, there are new developments in the epic battle during the discovery phase.
OpenAI suffered a significant legal setback when a judge ordered the company to disclose internal communications about deleting two massive pirated book datasets. The court ruled OpenAI waived attorney-client privilege by inconsistently claiming the datasets were removed due to “non-use,” then trying to shield deletion details.
This discovery could prove willful infringement, potentially increasing damages to $150,000 per work and affecting billions in liability. The ruling strengthens authors’ arguments that downloading pirated materials itself constitutes infringement, regardless of usage. OpenAI’s legal team faces depositions. If evidence destruction is proven, juries may assume deleted materials were unfavorable to the company.
The Hollywood Reporter – Winston Cho – November 26, 2025
These cases underscore the ongoing legal challenges and lack of definitive court decisions regarding “fair use” in AI training. Many cases will be moving toward summary judgment hearings scheduled for 2026.