Adobe Faces Class-Action Lawsuit Over AI Training
-

Adobe is being sued over allegations that its SlimLM AI model was trained using pirated books, including works by author Elizabeth Lyon.
The lawsuit claims Adobe used a derivative of the RedPajama/Books3 dataset, which allegedly included copyrighted works without consent.
SlimLM, Adobe’s small language model for document assistance, was pre-trained on SlimPajama-627B, released by Cerebras in 2023.
Similar cases have targeted Apple, Salesforce, and Anthropic, with Anthropic recently agreeing to $1.5B settlements over AI training datasets.
This case highlights the growing legal scrutiny over copyright and AI training data.