Baltimore has taken Elon Musk’s xAI to court, becoming the first U.S. city to sue over Grok’s wave of sexualized deepfakes. In a new complaint, city lawyers say the chatbot pumped nonconsensual intimate images into residents’ feeds and, in some instances, generated visuals that local investigators concluded met federal definitions of child sexual-abuse material. Baltimore is asking a state judge for injunctions that would force changes to Grok and X, along with the maximum penalties allowed under Maryland consumer-protection law.
According to CNBC, the city accuses xAI and X of deceptive and unfair trade practices and wants the court to order the companies to “cease the targeting and exploitation of Baltimore’s residents,” overhaul platform design, and revise their marketing. The filing says Baltimore users were exposed to a stream of manipulative, sexualized imagery generated by Grok that the city links to real-world harms.
The case lands on top of other growing legal trouble for Grok. Earlier this month, three Tennessee high-school students filed a proposed class action alleging the system morphed their photos into sexually explicit images, according to AP. The complaint in that case details images investigators said meet the federal statutory definition of CSAM and is available in the public docket at Ars Technica.
What Baltimore Wants From The Court
In addition to civil penalties, Baltimore is asking the court to force technical and policy changes that would limit Grok’s ability to generate and surface nonconsensual intimate images, and to rein in how X amplifies outputs from the chatbot. The city also wants reforms to how the company markets Grok and the platform so that users are clearly warned about the risks of having their photos ingested and repurposed by the model, according to CNBC. The complaint asks for the maximum civil penalties permitted and for an order blocking what Baltimore describes as deceptive practices.
The lawsuit lands amid mounting international alarm over Grok’s image tools. A meeting report from the Council of Europe citing Internet Watch Foundation data highlights a sharp rise in AI-related child sexual-abuse reports, noting that 97% of the IWF’s 2024 reports involved girls, a dynamic that advocates argue makes nonconsensual AI imagery a distinctly gendered threat. Regulators and civil-society groups in Europe and elsewhere have opened probes or requested detailed technical and policy information from xAI about how Grok operates.
Why SpaceX Is Suddenly In The Picture
Enforcement is further complicated by a recent corporate shuffle. In early February, xAI was folded into SpaceX, a move that observers say expanded the operational and financial muscle behind Grok. The Los Angeles Times reported on the transaction and on plans to pair AI models with large-scale compute tied to SpaceX’s satellite and launch business. That consolidation could shape where responsibility and data control ultimately sit if courts order technical fixes or ongoing oversight.
Legal Implications
Baltimore’s complaint leans on state consumer-protection law to seek injunctions and civil penalties, while parallel civil suits raise criminal issues where outputs allegedly meet federal CSAM definitions. The California filing in Doe v. xAI explicitly claims some Grok outputs qualify as child sexual-abuse material under 18 U.S.C. § 2256, a point laid out in a filing published by Ars Technica, which could spur criminal investigations and complicate platform defenses that rely on immunity doctrines for third-party content.
For people in Baltimore, the suit marks a new front in the push to hold AI developers and platforms to account for harms that start online and keep spreading offline. Other cities and states are likely to watch closely as courts weigh consumer-protection tools, evidence tied to potential criminal imagery, and the technical realities of how generative models are built and deployed.
