Elon Musk's AI company, xAI, is pushing to reveal the identities of four individuals who allegedly had deepfake sexualized images created of them using its Grok chatbot, including one reportedly depicted as a child. The move comes as part of an ongoing class-action lawsuit against xAI.
Filed in federal court, the lawsuit names the plaintiffs as South Carolina Doe, South Carolina Roe, New Jersey Doe, and Ohio Doe. These individuals have detailed the severe emotional distress caused by the alleged deepfakes, created earlier this year. They fear further online harassment and doxing if their real names are exposed in the litigation against xAI.
Lawyers for the plaintiffs argue that xAI's request to strip them of their anonymity is a tactic to intimidate them into dropping the case. They state that xAI is attempting to compound the harm already inflicted by the alleged deepfakes. The firm representing the individuals noted that xAI is asking the court to reverse its previous decision that allowed for pseudonyms, suggesting the company downplays the severity of the issue.
The controversy surrounding Grok erupted in January when users reportedly generated fake undressed and sexualized images of women, including apparent children, on the X platform. An analysis suggested Grok was used to create millions of such images in a short period. SpaceX, which now owns xAI, has reportedly set aside over $500 million to address the fallout from these issues and related lawsuits.
Initially filed with a single pseudonymous plaintiff, the case was refiled in May with the four main claimants. While the plaintiffs are willing to provide their personal information to xAI privately, they wish to maintain anonymity in public to protect their privacy and prevent further online harassment. The deepfake images themselves have not been included in any public court filings.
xAI's legal team contends that civil court rules generally require parties to be publicly identified, citing a public interest in knowing who is suing the company. They argue that no evidence has been presented to show specific further harm or threats to the individuals and that since the deepfakes will remain sealed, privacy concerns should be minimal. They claim that the mere fact a deepfake was created, without revealing the image, is not inherently stigmatizing.
Legal experts suggest that forcing plaintiffs in privacy cases to use their real names can deter litigation, creating an unjust situation where powerful entities might intimidate victims into silence. The plaintiffs have stated they would withdraw from the proceedings if forced to reveal their identities, emphasizing that the case involves highly personal and embarrassing deepfakes disseminated without their consent.
One plaintiff described discovering a deepfake image of themselves in revealing attire, causing significant worry about professional repercussions and further online targeting. They fear retaliation from supporters of Elon Musk and his companies, potentially leading to more extreme deepfakes. Another plaintiff, identified as New Jersey Doe, discovered deepfakes of himself after requesting Grok not create images of him without consent, believing this request drew the attention of online trolls.
South Carolina Roe alleges that Grok was used to create explicit images of her as a child, involving her father and criminal charges related to child sexual abuse material. She expressed distress over the lack of control and the fear of the images being shared online by predators, stating that public identification would cause "untold harm."
