Adobe hit with proposed class-action, accused of misusing authors’ work in AI training (techcrunch.com)

Adobe is facing a proposed class-action lawsuit from author Elizabeth Lyon, who claims the company improperly used pirated books, including her own, to train its AI model, SlimLM. The lawsuit alleges that Adobe's SlimPajama dataset, which was used for pre-training SlimLM, is a derivative of the controversial RedPajama dataset—a compilation of copyrighted works that has been the center of various legal disputes in the tech industry. Lyon contends that her copyrighted writings were included in this dataset without her consent, highlighting ongoing concerns regarding the ethical use of proprietary content in AI training. This lawsuit is significant for the AI/ML community as it underscores the growing legal scrutiny surrounding AI training datasets and the use of copyrighted material. With AI algorithms increasingly relying on vast, often unregulated datasets, such allegations may prompt greater accountability among tech firms regarding intellectual property rights. This case follows a recent trend where companies like Anthropic have faced hefty financial settlements for similar claims, indicating a potential shift in how content creators' rights are protected in the evolving landscape of AI technologies. As the industry continues to grapple with these issues, the outcomes of such lawsuits could shape future practices around AI training data.