UK MP Jess Asato Sues xAI Over Grok Deepfake Design Choices

Labour MP Jess Asato filed a High Court claim against Elon Musk’s xAI on June 3, 2026, seeking damages over non-consensual sexualized deepfakes that Grok’s image-editing tool generated of her, and by Thursday new claimants had already contacted her lawyers. The suit names xAI’s design choices as the source of harm, making it the first test in England of whether an AI developer can be held liable for how it built its product.

AWO, the data rights firm handling the case, confirmed it was already representing multiple individuals hoping to take action against xAI before the week ended. Many had tried to persuade X to remove the images on their own and gotten nowhere until legal support arrived, according to Ravi Naik, legal director of AWO.

What Grok Generated in 11 Days

Elon Musk announced the image-editing feature as a one-click capability on December 29, 2025. Users on X could respond to any posted image with a Grok prompt, and the chatbot would return a modified version. Within hours of launch, prompts generating results included “put her in a bikini” and “remove her clothes.”

The Center for Countering Digital Hate’s research into Grok’s image-generation volume (CCDH), a nonprofit study based on a random sample of 20,000 images generated between December 29 and January 8, produced these estimates:

• ~3 million sexualized images generated in the 11-day window
• ~23,000 of those appeared to depict children
• 190 images per minute, with one sexualized image of a child produced roughly every 41 seconds
• 4.6 million total images Grok’s editing feature generated across the period

xAI moved on January 9 to restrict image generation to paid subscribers and added technical blocks on prompts designed to undress subjects on January 14. Neither measure held entirely. Reuters found in early February that Grok was still generating sexualized images of subjects who explicitly had not consented. NBC News reported the same in April. Independent deepfakes researcher Genevieve Oh called Grok “the largest nonconsensual synthetic nudity generator” in the world in April, estimating its output likely surpassed all other nudifier tools combined.

The CCDH study noted that 29 percent of the sexualized images of children identified in its sample remained accessible on X as of January 15, days after xAI said it had acted.

Liability at the Design Level

The Architect Standard

At a press briefing Thursday, Ravi Naik framed the suit’s core reasoning plainly. “Just as if you’re an architect and build a building, you have liability for that architecture,” he said. “Those that build and deploy AI models make design choices about how these models operate. This will be the case that looks at liability for decisions in those design choices.”

The claim, set out in AWO’s official announcement of the UK’s first High Court action against Grok’s deepfake output, invokes the UK’s Data Protection Act and the tort of misuse of private information. AWO’s position is that xAI processed Asato’s personal data, her face and likeness, when Grok generated and distributed content depicting her without her knowledge or consent.

Grok created deepfake pornography and sexualised content which harmed thousands of women and children. Its ability is not an accident, nor misuse, it is a design choice by its creators. In launching this case, I am pursuing accountability for those choices.

Asato, Labour MP for Lowestoft, released that statement as she announced the High Court filing. Naik described the content in the same briefing as “built deliberately,” adding that it “existed because of design choices made by xAI.”

What the Claim Asks For

Asato is pursuing three remedies: financial damages, a formal judicial finding that xAI’s conduct was unlawful, and an injunction requiring xAI to cease ongoing violations and introduce safeguards the company chose not to build at launch. The injunction request carries the most structural weight, asking a UK court to retroactively impose design standards on a deployed commercial AI system.

To the BBC, she framed it with a car analogy. “It matters that the car was produced with the fault in the first place,” she said, arguing xAI’s January restrictions do not dissolve the harm already caused. Counsel on the claim includes Marie Demetriou KC of Brick Court Chambers and Edward Craven KC and Rosalind Comyn of Matrix Chambers.

New Claimants and the Case’s Scope

Asato announced the filing and called on others subjected to “abusive or demeaning” Grok-generated content to come forward. A handful contacted her lawyers the same day. AWO opened an email address for potential new claimants, noting it could not offer immediate or non-legal support and directing those in distress to resources at police.uk.

Asato told reporters she found the experience “psychologically distressing.” “This goes to the core of understanding what it means not to consent to something which literally strips your clothes off and makes you vulnerable,” she said. After she publicly criticized Grok in January, a video depicting her being chloroformed and prepared for sexual assault circulated on X in response to a post Musk had amplified. “Musk actually amplified the hatred against me,” she said, “which then led to the video that really was horrific.” On the same day she announced the June proceedings, a new AI-generated image of her in a bikini appeared on X, produced using a different tool.

Keir Starmer called Asato “absolutely right” and described the images as “disgusting.” Peter Kyle, the business secretary, said UK politicians needed to be “assertive” in holding Musk to account, describing him as “a complex and extreme person” who had taken “a much more active and extreme role in British politics.”

Asato said she wanted the legal action to establish that “AI companies are responsible for the design choices that they make when they launch their products” and to “remind them that they cannot act with impunity.”

The Cases Piling Up Against xAI

Asato’s High Court claim joins a queue of legal actions that has grown steadily since January:

1. January 2026: California’s attorney general issues a cease-and-desist to xAI; the European Commission opens a formal investigation under the Digital Services Act (DSA), the EU’s content liability framework for large platforms.
2. March 16, 2026: Three Tennessee teenagers file a class action in California alleging Grok produced child sexual abuse material from their school and family photographs via a licensed third-party app.
3. March 24, 2026: Baltimore becomes the first U.S. municipality to sue xAI, claiming Grok violated consumer protection law.
4. March 26, 2026: Amsterdam’s District Court issues a preliminary injunction, the first binding European court order against an AI image generator, prohibiting xAI from generating non-consensual sexualized imagery in the Netherlands, with daily fines of €100,000 per defendant capped at €10 million each.
5. June 3, 2026: Asato files at the High Court in London.

The Amsterdam ruling carries weight beyond Dutch borders. The court held that xAI, as designer and operator of Grok, bore responsibility for preventing unlawful outputs, regardless of who issued the prompts. That is precisely the reasoning AWO is advancing at the High Court.

On March 9, the same day xAI’s lawyers submitted a written statement claiming Grok could no longer generate non-consensual sexualized images, a courtroom demonstration in Amsterdam produced exactly that content from a single uploaded photograph. The court found the categorical denial irreconcilable with that evidence.

UK communications regulator Ofcom has opened a formal investigation under the Online Safety Act; Ireland’s Data Protection Commission launched a GDPR (General Data Protection Regulation) probe in February; and the California attorney general’s cease-and-desist remains active alongside the EU Commission’s DSA proceedings.

On SpaceX’s Books

xAI merged into SpaceX in February 2026, placing every Grok-related liability on the balance sheet of a company now preparing for a public offering expected later this month.

The SpaceX UK retail offer disclosure on file with the SEC shows the litigation exposure numerically. As of December 31, 2025, SpaceX had accrued $530 million for probable and reasonably estimable litigation losses. By March 31, 2026, following settlements and payments, that reserve had drawn down to $399 million. The document lists AI-related proceedings among significant matters and notes that additional losses beyond what has been accrued “may be material and are generally not reasonably estimable.”

Reports during the offering process noted that banks and advisers working on the deal were encouraged to subscribe to Grok and integrate its AI tools into their operations, a disclosure dynamic that drew scrutiny given the same product generated the litigation exposure now on their prospective issuer’s books.

Marex Financial, reviewing SpaceX’s disclosures for the UK retail tranche, noted its assessment “did not identify undisclosed litigation or regulatory matters that would prevent the offer from being distributed.” Its review covered disclosed matters; the Asato filing postdated that review.

SpaceX’s IPO, priced at $135 per share, is expected later this month. Grok’s litigation exposure appears in the prospectus risk factors by name.

Frequently Asked Questions

How Can People Affected by Grok’s Deepfake Feature Get Legal Support?

AWO has set up grokclaims@awo.agency for anyone subjected to non-consensual Grok-generated content who wants legal support or to join potential group action. The firm says it cannot offer immediate or non-legal support, and directs those in immediate distress to resources available at police.uk.

Is Creating a Non-Consensual Deepfake Illegal in the UK?

Yes. Legislation passed in 2025 makes it illegal to create or request a non-consensual deepfake image of an adult in England and Wales. Asato’s case additionally invokes the Data Protection Act and the tort of misuse of private information, both of which are independent of the deepfake statute and target the developer of the platform rather than the individual who submitted a prompt.

How Does the Asato Claim Differ from the Baltimore Lawsuit?

Baltimore’s March 2026 suit relies on the city’s consumer protection statute, arguing xAI marketed Grok as safe while knowing otherwise. AWO’s High Court claim uses UK data protection law and the privacy tort of misuse of private information, centered on xAI’s processing of Asato’s personal data without consent. Both suits hold that the source of liability is xAI’s product design decisions as the system’s developer.

Has xAI Responded to Asato’s High Court Filing?

xAI had not publicly commented on the claim as of the time of publication. The company did not respond to requests for comment.

What Happens Next for the Case in the High Court?

After filing, proceedings are served on xAI, which then has a defined period to file a defense. A case management hearing follows to set the litigation timetable. No trial date had been announced as of June 7, 2026.