AINEWS
A federal class action leans on a 2025 Tennessee statute to argue the chatbot and image tool companies had a duty to install safeguards before that harm occurred.
A federal class action filed July 7 in Tennessee on behalf of five anonymous child plaintiffs, three of them from Tennessee, accuses xAI, the Elon Musk-backed maker of the Grok chatbot, and Stability AI, the image-model company, of enabling the creation of AI-generated child sexual abuse material from the children's photographs without consent. The amended complaint argues that the companies had a duty to install industry-standard safeguards before that harm occurred, and that they failed to do so. The docket for Doe 1 v. xAI Corp. is public on CourtListener.
The lawsuit runs on two legal tracks. The first is a federal claim that the companies owed a duty of care to prevent the generation of child sexual abuse material through their tools, and that they breached it. The second is a state-law claim under a Tennessee statute enacted in 2025 that creates criminal penalties for technology designed or marketed to produce AI-generated child sexual abuse material. The plaintiffs argue that the statute reaches general-purpose AI platforms, not only narrow tools built for that purpose.
The complaint specifically points to Grok's optional "Spicy Mode," a feature plaintiffs describe as offering more adult-oriented content. According to Ars Technica's review of the filing, a single Grok user allegedly generated roughly 7,000 child-sex images before the platform reported the activity, and reported only one prompt to law enforcement, a gang-rape scenario, before that point. Lieff Cabraser Heimann & Bernstein, the plaintiffs' firm, has framed the case as a test of platform accountability for downstream misuse of image-generation tools.
Both xAI and Stability AI deny the allegations, and the case remains pending. No court has ruled on the merits. Plaintiffs' counsel Mark Chalos, managing partner of Lieff Cabraser Heimann & Bernstein's Nashville office, said the alleged conduct is "a widespread problem" the companies "have known about for a very long time."
A ruling for the plaintiffs on the federal duty-of-care theory would establish that general-purpose AI companies can be held liable when their tools are used to generate child sexual abuse material, even when the company did not design the tool for that purpose. A ruling for the defendants, treating the tools as neutral platforms rather than designers of a foreseeable misuse, would narrow the suit's precedential value and put similar cases elsewhere on shakier ground.
The 2025 law predates any major judicial test of how state AI-CSAM statutes apply to consumer chatbots and image generators, and this case would be the first. A ruling for the plaintiffs would make Tennessee an early proving ground for state-level AI safety enforcement, with other states likely to study the reasoning. A ruling for the defendants would push the question back to federal law, where it would sit alongside Section 230 and product-liability defenses the companies have signaled they intend to raise.
Prior CSAM litigation against generative-AI platforms has foundered on whether the company knowingly enabled the harm or merely provided a tool that some users abused. The Doe 1 v. xAI complaint tries to thread that needle by tying the alleged safeguards gap to specific platform features, and by adding the Tennessee statute as a separate hook that does not require proving the companies intended the result.
Watch items: any motion to dismiss from xAI or Stability AI on Section 230 or product-liability grounds; any guidance from Tennessee Attorney General Jonathan Skrmetti on how the 2025 statute should reach general-purpose AI; and the first substantive ruling on either the federal duty-of-care theory or the state statute. A motion to dismiss is the most likely next step, and the court's reasoning there will signal how far the case can run.