In short
- The complaint estimates that Grok created 3 million sexually explicit images in a few days, including thousands featuring children.
- Baltimore is seeking civil penalties, restitution to the affected citizens, and a court order to stop what it says is happening.
- Legal experts say the case will depend on whether the courts view AI systems as harmful agents rather than tools.
A Baltimore lawsuit against Elon Musk’s xAI and his Grok chatbot could decide how far cities can go to control artificial intelligence without federal law, according to one expert.
The Mayor and City Council of Baltimore said the defendant X Corp., xAI, and SpaceX in a Maryland court, claiming that the companies violated local consumer protection laws by creating and deploying Grok, an AI system known to create and distribute sexually inappropriate images, including children.
The lawsuit alleges that Grok enables users to “undress” or “edit images” of real people with little to no coercion, exposing citizens to privacy violations and emotional harm, according to the lawsuit. words from the DiCello Levitt law firm, which is representing the city along with the Baltimore City Law Department.
“These crimes, especially those involving children, have painful, lifelong consequences for the victims,” Baltimore Mayor Brandon M. Scott said in a statement.
The case led to Grok’s international investigation, including a nationwide investigation US, EU, France, UK, Australiaand Irelandand the federal portion that was issued last week is three Tennessee kids protest the tool created interactive content for children using their real photos.
“This case can be seen as a way for the city to regulate AI in the absence of federal laws, using consumer protection and public harm doctrines to keep AI companies established,” said Ishita Sharma, managing partner at Fathom Legal. Decrypt.
“In general, while users who are causing harm will be part of the dispute, strong laws should be if the AI system itself has contributed,” said Sharma, adding that if the courts see Grok “as an active creator instead of just an intermediary,” the responsibility will fall more on xAI.
By design
The suit in Baltimore alleges that the company “manufactured, sold, and distributed” Grok, knowing it could produce inappropriate images that resembled child sexual abuse, while publicly stating that the products were illegal.
The complaint alleges that Grok made between 1.8 million and 3 million sexually explicit images in a few days between December 29, 2025, and January 8, 2026, inclusive about 23,000 show childrenAccording to the Center for Counting Digital Hate and a New York Times analysis.
Baltimore says the operation began when Elon Musk developed Grok’s appearance answer “Perfect” to the photo of her bikini created by the tool, which rises from about 300,000 pictures in the nine days before posting to about 600,000 a day on X.
“X is now one of the largest distributors of NCII and CSAM,” the lawsuit reads, citing the platform’s policy that prosecutors block against content that constitutes false testimony.
“Evidence of defense delays or inaction in the face of perceived dangers would strengthen claims of negligence or recklessness,” Sharma said, adding that dismissal of the case appears unlikely, ruling out the possibility, although the case could lead to an “AI-based verdict.”
The city is seeking civil penalties, injunctive relief, restitution, and disgorgement of profits.
Decrypt reached out to Musk via xAI and SpaceX for comment.
Daily Debrief A letter
Start each day with top stories right here, including originals, podcasts, videos and more.
