Adobe AI Training Lawsuit Intensifies as Authors File Proposed Class Action

In Focus

• Adobe class action lawsuit alleges unauthorized use of copyrighted books to train AI models
• Plaintiff claims Adobe misused authors’ work via datasets connected to SlimLM training
• Lawsuit adds to mounting Adobe AI Training Lawsuit trend hitting major tech players
• Potential industry implications for AI copyright infringement standards

On December 17, 2025, a new Adobe class action lawsuit was filed accusing Adobe Inc. of misusing copyrighted books, including works by Oregon-based author Elizabeth Lyon, in its AI training process, according to TechCrunch.

The complaint, filed in the U.S. District Court for the Northern District of California, asserts that Adobe’s AI development relied on material authors did not authorize for such use.

Legal Challenge Targets Adobe’s AI Training Practices

In the suit filed as Lyon v. Adobe Inc., Lyon alleges that Adobe incorporated her instructional books and those of other copyright holders into the training data for its SlimLM small language model without consent. The complaint claims Adobe relied on SlimPajama-627B, a derivative of the RedPajama dataset, which in turn includes the Books3 corpus believed to contain roughly 191,000 copyrighted books.

The lawsuit focuses on alleged misuse of copyrighted material in AI model training, positioning itself within broader scrutiny of tech companies’ data practices. According to Lyon’s complaint, the chain of datasets leading from Books3 into Adobe’s AI infrastructure represents an unauthorized use of literary works for commercial AI development.

The complaint seeks unspecified monetary damages on behalf of a proposed class of all copyright owners whose works were allegedly included without authorization. Should the court grant class certification, the litigation could broaden the number of affected authors far beyond Lyon herself, raising potential exposure for Adobe.

Key details from this lawsuit include

• Filing: Lyon v. Adobe Inc., Northern District of California federal court.
• Plaintiff: Elizabeth Lyon, author.
• Defendant: Adobe Inc.
• Allegations: Unauthorized use of copyrighted books in AI training.

Book Authors and Broader Industry Precedent

The Adobe class action lawsuit follows a wave of similar disputes alleging copyright infringement in AI training practices. Earlier actions against other AI developers, including a notable $1.5 billion settlement with Anthropic, illustrate how authors are increasingly asserting legal claims over the use of copyrighted works in generative AI systems. In other news, Adobe has deepened its investment in multi-model generative AI by integrating Gemini 3 Nano Banana Pro into Adobe Firefly and Photoshop, offering an expanded set of AI image generation and editing capabilities.

“The SlimPajama dataset was created by copying and manipulating the RedPajama dataset (including copying Books3)… Thus, because it is a derivative copy of the RedPajama dataset, SlimPajama contains the Books3 dataset, including the copyrighted works of Plaintiff and the Class members,” Lyon’s lawsuit stated, as per TechCrunch.

Future of AI Training Standards

The Adobe AI Training Lawsuit underscores the evolving legal environment surrounding generative AI. Businesses that develop, license, or integrate AI systems should take note of how copyright law is being applied to foundational AI training datasets. A ruling for the plaintiffs may influence how companies document dataset sources and negotiate rights with content owners. Recently, Adobe has acquired SEMrush to boost its generative AI product portfolio.

Should the court grant class certification and find for the plaintiffs, the repercussions could extend beyond Adobe to other AI providers, especially those utilizing broad scraped datasets. This Adobe class action lawsuit is not an isolated event but part of a trend where authors and rights holders are actively challenging prevailing AI development norms.