AI-assisted artwork poses a simple question: When can an artist using AI tools copyright their work? Early this year, the Copyright Office addressed this issue and rejected the proposition that only prompting an AI model can create a copyrightable work. But their analysis missed that “randomness” for a computer means something entirely different than we generally think, ultimately underselling the amount of control someone can have over the model’s output.
Copyright Law Year in Review, Dec. 11, 2025
A computer scientist is asking the U.S. Supreme Court to extend copyright protection to works created entirely by artificial intelligence in one of the first cases to reach the justices about the revolutionary technology.
The U.S. Supreme Court is set to commence its term on October 6. Among the cases it will review are several appeals concerning copyright and trademark law. One notable case seeks to address procedural inconsistencies and claims of unconstitutional vagueness attributed to the U.S. Court of Appeals for the Federal Circuit.
A federal judge in the Northern District of California granted preliminary approval to a $1.5 billion settlement between Anthropic and a class of book authors who alleged that the artificial intelligence company used their copyrighted works to train its chatbot Claude without their consent. The settlement is the largest copyright settlement of all time, covering 482,460 works and paying authors slightly more than $3,000 per work infringed.
A new lawsuit against Perplexity AI claims responses generated by the artificial intelligence platform violate the trademarks of Encyclopaedia Britannica and Merriam-Webster by attributing false information to their widely esteemed brands. The complaint alleges Perplexity’s generative AI “answer engine” violates the plaintiffs’ copyrights and also cites them as sources of false or incomplete information.
A federal judge in the Northern District of California granted preliminary approval on September 25 to a $1.5 billion settlement between Anthropic and a class of authors who alleged that the artificial intelligence company used their copyrighted works to train its chatbot Claude without their consent. The settlement is the largest copyright settlement of all time, covering 482,460 works and paying authors slightly more than $3,000 per work infringed.
How are copyright holders to prove their works were used to train AI models if the details about the vast data sets used for such training are kept secret? That’s a dilemma that surfaced in late August when a federal judge dismissed a claim of direct infringement raised by a group of authors.
How are copyright holders to prove their works were used to train AI models if the details about the vast data sets used for such training are kept secret? That dilemma surfaced when a California federal judge recently dismissed a claim of direct infringement raised by a group of authors.
In late July, two important decisions came down from courts in the Northern District of California regarding the unauthorized use of copyrighted material for the training of large language models. No real consensus has emerged as to the effect they will have on the broader AI litigation landscape.
© Copyright 2026 ALM Global, LLC trading as Centellic.
Advertise with Us Customer Support Terms of Service Privacy Policy
