Getting your
Trinity Audioplayer ready...A federal judge in San Jose has rejected an attempt by Elon Musk’s artificial intelligence company xAI to publicly identify the four people suing the company over sexualized deepfakes allegedly created by its Grok chatbot, ruling that revealing their names could expose them to further harassment and fabricated images.
Judge Casey Pitts said Thursday that the three women and one man had shown anonymity was warranted because of the “sensitive and highly personal claims” in the case and the threat of retaliation against people who challenge the creation and distribution of deepfakes.
Publicly identifying them could cause “severe” harm, Pitts wrote, including emotional distress and damage to their reputations and employment if additional sexualized images were created and shared.
The ruling came over the objections of xAI, which argued that allowing the plaintiffs to remain anonymous would impede its ability to find witnesses and violate the public’s right of access to court proceedings.
The plaintiffs are seeking class-action status on behalf of what their lawsuit claims are hundreds of people whose photographs Grok altered to depict them nude, in sexual positions or in explicit videos. One plaintiff alleges the chatbot generated an image suggesting she had sex with her father.
Grok, which is built into the social media platform X and available as a standalone app, can generate “deepfakes” — seemingly realistic but fabricated photos or videos — in response to users’ prompts.
The lawsuit, filed in January in San Jose U.S. District Court, claims Grok “humiliates and sexually exploits women, girls, and others by altering real photos of them so that they are undressed and/or posed in sexual positions or featured in explicit videos.”
It also alleges Grok produced “illegal child sex abuse material.”
In May, xAI asked the court to reverse an earlier ruling allowing the plaintiffs to remain anonymous while the case proceeds. The company argued that there was “nothing inherently stigmatizing to (the plaintiffs) about the fact that a third party created an undisclosed deepfake image of them.”
The company also disputed what it called “speculative allegations of potential future harm caused by the further dissemination of deepfake images online.”
Pitts rejected those arguments Thursday, noting that the lawsuit alleges “individuals who have opposed the creation and distribution of deepfakes are commonly subject to vicious campaigns of harassment and intimidation.”
If their anonymity were lifted, the plaintiffs “would be vulnerable to the creation and distribution of further deepfakes in retaliation” for filing their lawsuit, Pitts wrote. One plaintiff, identified as “South Carolina Roe,” alleges in the lawsuit that a police officer awakened her at the home she shared with her parents in February and told her that her father faced criminal charges over the alleged possession of child sexual abuse material.
“South Carolina Roe has learned that Grok altered several images depicting South Carolina Roe in a bed with her father,” the lawsuit claims. “Grok altered the images to make it appear as if they had just consummated a sexual act.”
Additional deepfakes of the woman may have been generated by Grok, but South Carolina Roe has “yet to find out about them because they are subject to an ongoing criminal investigation,” the lawsuit said.
Another plaintiff, “South Carolina Doe,” posted a photograph of herself fully clothed on X on Jan. 2, according to the lawsuit.
The next morning, she discovered that Grok had used the photograph to create and publicly post an image depicting her in a revealing bikini, the lawsuit alleges.
“She was shocked and embarrassed by the deepfake, and it caused her to panic as she was overwhelmed with thoughts of who would see the deepfake and think that she had taken the image herself,” the lawsuit said.
She worried that her employer or co-workers could see the image and that it could damage her career, according to the lawsuit.
“She was overcome with disgust at the thought of what the X user who had asked Grok to create the deepfake of her was doing with the photo,” it said.
South Carolina Doe began reporting the image to X almost immediately, but the company initially refused to remove it, the lawsuit alleges. She also complained directly to Grok through X.
“Grok denied creating the deepfake, denied posting any images since January 1, 2026, and claimed it did not have image generation or editing capabilities,” the lawsuit claims.
X removed the image three days later, according to the lawsuit.
The male plaintiff, identified as “New Jersey Doe,” alleges that Grok altered a shirtless photograph he had posted while wearing pants, creating and publishing versions that depicted him in a jockstrap and a thong. X refused his requests to remove the images, the lawsuit claims.
The fourth plaintiff, “Ohio Doe,” alleges that Grok altered a photograph of her fully dressed to depict her wearing a bikini made from plastic wrap and, in one image, kissing a man.
“It took months,” the lawsuit said, “for the deepfakes to be taken down.”
The rollout of Grok’s “edit image” function in December prompted widespread reports that people were using the chatbot to create child sexual abuse material. That led to international backlash, lawsuits and investigations, including one by the State of California. State officials did not immediately respond Friday to questions about the investigation’s status.
xAI did not respond Friday to requests for comment and has not filed a response directly addressing the lawsuit’s allegations. Musk is not named as a defendant, but the lawsuit claims he and xAI “have chosen to capitalize on predators’ insatiable appetite for non-consensual sexual images and videos of real people, including children.”
In a separate lawsuit filed July 27, xAI challenged Minnesota’s ban on apps and websites used to digitally undress people, arguing that the law is overly broad and unfairly penalizes companies. xAI said in that lawsuit that its terms of service and usage policies prohibit using Grok to undress people or depict them sexually.
Last month, xAI sued a Texas man after his arrest on suspicion of sexually exploiting a minor. The company alleged that the man used Grok to create child sexual abuse material.