Mayor and City Council of Baltimore v. X Corp

Issue: In *Mayor and City Council of Baltimore v. X Corp.*, the City of Baltimore argues that X Corp. and affiliated entities violated Baltimore's Consumer Protection Ordinance by publishing safety policies that expressly prohibited non-consensual intimate imagery and child sexual abuse material while simultaneously operating a generative AI system — Grok — that produced millions of such images, including approximately 23,000 depicting minors, during an eleven-day period in January 2026. The central legal questions are whether Grok's autonomous image output constitutes the defendants' own content creation rather than third-party content (thereby defeating Section 230 immunity), and whether defendants' published acceptable-use policies were actionable false commercial representations under consumer-protection law.

The Mayor and City Council of Baltimore, acting through City Solicitor Ebony M. Thompson, filed this complaint in Baltimore City Circuit Court; the action was subsequently removed to the District of Maryland as Case 1:26-cv-02103-ELH. The complaint is an initiating pleading at the pre-answer stage, naming X Corp., x.AI Corp., x.AI LLC, and SpaceX as defendants. Plaintiff alleges that defendants' own policies — prohibiting deepfakes, non-consensual intimate imagery, and child sexual abuse material — were false when made and throughout the relevant period, because Grok was concurrently generating the very content those policies forbade. The complaint further alleges that when public criticism emerged, defendants restricted the image-editing feature to paying subscribers rather than disabling it, which plaintiff characterizes as monetizing a harmful tool rather than remedying it. Relief sought includes declaratory judgment, a permanent injunction, disgorgement, civil penalties, and attorneys' fees.

This complaint is among the first municipal consumer-protection enforcement actions to directly challenge a generative AI system's design as the source of harmful content, rather than targeting user-generated material hosted on a platform — a framing strategically constructed to route around Section 230 immunity. If courts credit the argument that a generative AI is itself an "information content provider" whose architecture, not user prompting, drives injurious output, the decision would meaningfully narrow the immunity that has historically insulated platform defendants from product-design liability. The policy-as-false-representation theory is the complaint's most doctrinally grounded pillar and could independently establish a template for municipal enforcement against AI companies whose published safety commitments diverge from actual system behavior. The inclusion of SpaceX based on an unconsummated acquisition, and the attribution of Elon Musk's personal social-media activity to corporate defendants, are legally thin theories that will test how far courts are willing to extend consumer-protection liability at the pleading stage in high-profile AI litigation.