IP Law Daily, COPYRIGHT—N.D. Cal.: Author files suit against Adobe alleging AI-based copyright infringement, (May 21, 2026)

By Joe Cox, J.D.

The suit, filed on the author’s own behalf, and on behalf of a class of similar authors, alleges that Adobe engaged in a “massive piracy scheme” to develop AI models.

On Monday, an author filed a copyright infringement suit on her own behalf and for a class of fellow authors in California federal court against Adobe Inc. for copyright infringement related to Adobe’s development of AI models. The suit alleges that Adobe engaged in a “massive piracy scheme” with apparently unlicensed use of millions of books to train AI models Nemotron and SlimLM. The complaint alleges that Adobe utilized shadow libraries that openly flaunt copyright law to obtain millions of books for AI development without valid licensing, and seeks to establish a class of similarly impacted authors and obtain relief from these violations. (Tanzer v. Adobe Inc., No. 3:26-cv-04712 (N.D. Cal. May 18, 2026)).

Allegations of the complaint. The named plaintiff, E. Molly Tanzer, is an author who resides in Colorado. She alleges that Adobe has partnered with NVIDIA to develop AI foundation models like Nemotron and SlimLM. In the absence of any indication that Adobe licensed the massive number of books which Tanzer indicated would be utilized in developing such models, Tanzer argues that the company has sourced “millions of copyrighted books” from pirated sources, including shadow libraries, without licensing payment or permission.

Tanzer specifically alleges that Adobe obtained pirated works through Anna’s Archive, one such shadow library, and through the Book3 dataset, a subset of a SlimPajama dataset which Tanzer alleges to include nearly 200,000 pirated books. Tanzer further notes that while Adobe apparently appropriately licenses images with payment to creators for its Firefly AI models, book authors have not been similarly compensated. Adobe has then, according to the complaint, incorporated AI models like Nemotron and SlimLM into its programs and tools, including Adobe Acrobat, and has profited greatly from same.

Tanzer details Adobe’s use of printed books to develop AI models, noting that Adobe officer Scott Belsky has claimed that Adobe only utilizes data that is “openly scraped,” a contention which Tanzer disputes. Tanzer details the training of Nemotron in particular as allegedly involving well over a million books. While she notes that Adobe and NVIDIA have not publicly revealed sources of books utilized, Tanzer alleges the use of shadow libraries, so named because they routinely flout copyright law, has occurred, and has thus allowed Adobe to avoid appropriate payment for copyrighted material.

Relief sought. Tanzer seeks to establish a class of similarly situated authors whose copyrights have been violated since at least May 18, 2023. She seeks certification of a class, judgment, damages, attorney fees, injunctive relief, and for Adobe to be responsible for the cost and expense of notifying class members.

The Case is No. 3:26-cv-04712.

News: Copyright CaliforniaNews AINews TechnologyInternet GCNNews