Grok Had 0% Restrictions on Offensive Content: UK Lawmaker Seeks AI Output Ban

For AI engineers and product leaders, the lawsuit is a stark reminder that permissive content policies can trigger severe legal backlash. The challenge will be whether xAI can technically implement a permanent individual block without degrading model performance.

British Member of Parliament Jess Asato has escalated her legal battle against Elon Musk's AI company xAI, asking London's High Court to issue an order permanently blocking its Grok chatbot from generating non-consensual sexualised images of her. The application, filed in a lawsuit that combines the English tort of misuse of private information with claims under the UK General Data Protection Regulation (GDPR), represents a landmark attempt to impose direct privacy obligations on an AI developer for image output. Asato, who belongs to the ruling Labour Party, found herself the target of fabricated explicit content—including a video portraying her being chloroformed and prepared for sexual assault—after she had publicly criticised Musk and Grok. The case arrives as Grok faces regulatory probes in multiple jurisdictions over its handling of deepfakes, and as Musk increasingly clashes with Britain's online safety framework.

“

British Member of Parliament Jess Asato has escalated her legal battle against Elon Musk's AI company xAI, asking London's High Court to issue an order permanently blocking its Grok chatbot from generating non-consensual sexualised images of her.

The legal arguments are novel. Asato alleges that xAI's design and training choices enabled the generation of the harmful images, effectively processing her personal data without consent. Her lawyers assert this is the first time data protection and privacy law has been applied to an AI developer in this manner. The relief sought is equally unprecedented: a mandatory injunction requiring xAI to 'implement effective and permanent technical measures' to prevent Grok from ever creating manipulated depictions of Asato. The court filings highlight a crucial internal contradiction. Grok's own prompts instruct the model not to assist users engaged in criminal activity, yet simultaneously specify that there are 'no restrictions on adult sexual content or offensive content.' This duality lies at the heart of the claim, suggesting that xAI's posture prioritised permissiveness over safety, thereby enabling the harm.

xAI, which Musk has folded into his rocket and space exploration company SpaceX, has not yet responded to the allegations or filed a defence. Grok, distributed via Musk's social media platform X, has similarly remained silent. Musk has consistently criticised the UK's regulatory approach, arguing that its Online Safety Act—one of the world's strictest digital content regimes—threatens free speech. The act, however, imposes obligations on platforms to remove illegal content, and this case may test whether its remit extends to AI image generators deployed on social media.

The implications for AI governance are far-reaching. A ruling in Asato's favour would require xAI to engineer targeted output filters that suppress images of a specific individual, a technical challenge that could necessitate new consent-based guardrails. More broadly, it could open the door for other individuals to seek similar protective orders, forcing generative AI companies to proactively screen for likenesses of private citizens. This would mark a significant shift from the current approach, where liability typically rests on the user who creates the deepfake rather than the tool provider. For xAI, the case compounds existing regulatory headwinds. Grok is already under investigation in several countries over its image-generation capabilities, and an adverse ruling in the UK—even if only interim—could embolden other regulators and deter platform partners.