In a move aimed at redefining legal accountability in the age of generative AI, xAI filed a lawsuit Tuesday against Terry Wayne Harwood. The firm accuses Harwood of using its Grok chatbot to generate illegal child sexual abuse material (CSAM), claiming the user deliberately bypassed safety safeguards.
The Harwood Case and Allegations
According to the complaint, Harwood—who was arrested earlier this year in South Carolina—utilized two xAI accounts between December 8 and February 18. xAI alleges the defendant attempted to "nudify" images of victims, including a young girl estimated to be as young as 10 years old. While Grok's safeguards rejected some prompts, Harwood reportedly used misleading language and modified prompts to circumvent content filters.
The "Neutral Tool" Defense
The lawsuit arrives as xAI faces a separate proposed class action representing victims who claim Grok was used to create thousands of sexualized images of minors. xAI's legal strategy is transparent: it wants the court to recognize Grok as a "neutral tool" entirely subject to user control.
The company argues that users agree to terms of service that explicitly prohibit sexualizing children or altering the likeness of real people. If the court upholds xAI’s indemnity clause, Harwood—and potentially any future bad actor—would be solely liable for damages resulting from AI outputs, shielding the platform from liability.
Legal Hurdles and Contradictions
However, this strategy faces significant hurdles. The US Copyright Office currently does not view AI outputs as human-created, which could complicate xAI's argument that the user is the sole creator of the content. Furthermore, a 2026 NCMEC report indicated that 90% of xAI’s CyberTipline reports lacked the actionable user information needed for law enforcement to track perpetrators, casting doubt on the firm's claims of proactive cooperation.
