IP Law Daily, COPYRIGHT NEWS: Class action complaint accuses Adobe of copyright infringement, (Dec 18, 2025)

By Steven D. Cole, J.D.

The proposed class action would hold the tech giant liable for the unauthorized use of copyrighted books to train its artificial intelligence software models.

An author who resides in Oregon, on behalf of herself and others similarly situated, has filed a class action complaint accusing Adobe Inc. (Adobe) of downloading, copying, storing, and using copyrighted books without authorization by the copyright holders in order to develop artificial intelligence (AI) software for commercial benefit. The complaint seeks class certification and relief in the form of statutory damages, actual damages, restitution of profits, a declaration of willful infringement, and other appropriate legal and equitable remedies (Lyon v. Adobe Inc., No. 5:25-cv-10732 (N.D. Cal. filed Dec. 16, 2025)).

Alleged infringement. Adobe is a computer software company that offers a wide range of programs. It is well known for its popular PDF viewers and editors. This particular controversy involves Adobe’s AI services.

According to the complaint, Adobe has promoted “SlimLM” as a part of its AI offerings and as a model intended for integration into mobile devices. SlimLM consists of a series of Adobe’s small language models (SLMs) that are optimized for document assistance tasks on mobile devices. SLMs are AI software programs designed to run on devices with limited hardware resources, such as smartphones and tablets, which reply to user prompts with natural-sounding text outputs (in contrast to large language models, which often run remotely with results transmitted via the internet to devices). Language models are developed by processing vast quantities of text from diverse sources (a “pre-training dataset”) as a means of learning statistical patterns and associations within that data. In order to develop a high-quality pre-training dataset, books and other texts are downloaded (or otherwise acquired and transmuted into digital form), copied, stored, and processed.

Adobe used a pre-training dataset called SlimPajama to develop SlimLM. SlimPajama is a cleaned-up version of the RedPajama dataset, which is comprised of public domain, licensed, and unlicensed copyrighted materials. The RedPajama dataset, in turn, contained a subset of books that was a copy of another dataset called “Books3,” which included the copyrighted works of the author and proposed class members. The essence of the complaint is that “Adobe downloaded and copied copyrighted works and the Infringed Works as contained in the SlimPajama dataset without authorization from, or after providing compensation to, their authors.” These acts are alleged to constitute direct copyright infringement under 17 U.S.C. § 501.

The complaint further alleges that Adobe has retained copies of these pre-training datasets, which contain copies of the purportedly infringed works, on its servers, and that Adobe continues to store and use them in training for new versions of SlimLM and other related models, as well as for undisclosed future uses.

Class certification. The complaint seeks Rule 23 certification for a class consisting of:

All persons or entities domiciled in the United States that own a United States copyright in any work that was downloaded, copied, stored, or used as training data by Defendant without authorization during the Class Period according to Defendant’s records.

The class period, as defined in the complaint, begins on December 16, 2022, and runs through the present, though the class period may be amended to comport with any evidence uncovered during further investigation or through discovery.

In furtherance of the request for class certification, the author has attested to the following: she has retained counsel with substantial experience in prosecuting complex litigation and class actions, and will fairly and adequately represent and protect the interests of the other members of the proposed class; common questions of law and fact predominate over any questions that may affect individual members of the proposed class; the class treatment of this controversy is superior to multiple individual actions or piecemeal litigation; Adobe has acted, and failed to act, on grounds generally applicable to the proposed class members; and the proposed class members have all suffered similar harm as a result of Adobe’s unlawful and wrongful conduct.

Requested relief. In addition to class certification and a demand for a trial by jury, the complaint explicitly requests a judgment against Adobe consisting of the following: an order declaring that Adobe’s conduct violates 17 U.S.C. § 501; an award of statutory and other damages under 17 U.S.C. § 504 for copyright violations; reasonable attorney fees and reimbursement of costs under 17 U.S.C. § 505 or otherwise; a declaration that such infringement is willful; destruction or other reasonable disposition of all copies of infringed works that Adobe made or used in violation of the exclusive rights the proposed class members, pursuant to 17 U.S.C. § 503(b); pre- and post-judgment interest on the damages awards; and any further relief that the court deems to be appropriate.

The Case is No. 5:25-cv-10732.

Judge: Cousins, N.

Attorneys: Yevgeniy Y. Turin (McGuire Law, P.C.) for Elizabeth Lyon.

Companies: Adobe Inc.

News: AINews Copyright TechnologyInternet CaliforniaNews