Tennessee Minors v. xAI (Grok CSAM Deepfake Class Action)

A perpetrator with a "close and friendly relationship" to at least one plaintiff gathered photos from their Instagram, yearbook, and social media accounts. Using a third-party application powered by xAI's Grok algorithm (not Grok directly on X), they generated sexually explicit deepfake images and videos of the three Tennessee teens — including one video of a plaintiff "undressing until she was entirely nude." The output appeared "entirely real."

Jane Doe 1 discovered the images around December 6, 2025, when an anonymous Instagram user directed her to a Discord server created by the perpetrator. On that server she found at least 5 sexually explicit files of herself and images of at least 18 other minor females. The perpetrator was arrested in December 2025.

The perpetrator also distributed CSAM on Telegram in group chats with "hundreds of other users," using the files as "bartering tools" to trade for other child sexual abuse material.

Jane Does 2 and 3 were notified by law enforcement around February 12, 2026 that their images had been found. One plaintiff's mother gave a public statement about her daughter experiencing panic attacks, saying her "senior year turned into a nightmare."

The lawsuit alleges xAI deliberately designed Grok's "Spicy Mode" for sexual content, configured system prompts to assume "good intent" for references to "teenage" or "girl," and that Musk personally promoted the "undressing" capability on X. The suit further alleges xAI "outsourced liability" by licensing its technology to third-party app makers "often outside the U.S.," generating additional profit while evading responsibility. After public exposure in January 2026, xAI restricted image generation to paid subscribers and blocked "undressing" — which the lawsuit frames as monetizing rather than preventing CSAM.

• Third-party application using xAI's Grok algorithm generated photorealistic sexually explicit images and videos of minors from ordinary social media photos
• Grok's "Spicy Mode" was allegedly designed to enable sexual content generation
• System prompts reportedly configured to assume "good intent" for queries referencing "teenage" or "girl"
• No effective CSAM prevention measures implemented despite industry-standard tools being available
• After exposure, capability was restricted to paid subscribers rather than removed — effectively monetizing it

xAI's Grok image generation technology, licensed to third-party applications, enabled the creation of photorealistic CSAM from ordinary social media photos of minors. The lack of CSAM prevention measures, combined with the licensing model that placed the technology in less-regulated third-party apps, created an ecosystem where a perpetrator could generate, distribute, and trade AI-generated child sexual abuse material at scale.

The psychological harm to victims includes panic attacks, fear, and severe distress from knowing sexually explicit images of themselves exist and have been widely distributed. The trading of these images as "bartering tools" for other CSAM compounds the exploitation.

• March 16, 2026: Class-action lawsuit filed in U.S. District Court, Northern District of California, San Jose Division by Lieff Cabraser Heimann & Bernstein and Baehr-Jones Law
• December 2025: Perpetrator arrested after Jane Doe 1 discovered images on Discord server
• February 12, 2026: Jane Does 2 and 3 notified by law enforcement

Proposed class includes all U.S. residents reasonably identifiable in sexualized Grok-generated images, with particular focus on minors. Legal claims include Masha's Law, Trafficking Victims Protection Act, product liability, negligence, public nuisance, defamation, IIED, California Right of Publicity, and California Unfair Competition Law.

xAI did not respond to requests for comment.