World Technology

An MP is suing over AI-generated images of her. The argument is about design, not users

The company's own defence says the tool is built to turn a short prompt into a deepfake. That admission is now the centre of the case.

An MP is suing over AI-generated images of her. The argument is about design, not users

The Grok chatbot was designed by Elon Musk's SpaceX Super Intelligence to generate sexualised and offensive images, an MP suing the company has said.

Jess Asato, Labour MP for Lowestoft in Suffolk, is suing SpaceXSI for damages after images and videos of her were generated and shared online using the AI tool. Her claim alleges the company's design choices make it responsible for Grok's output, and that it misused her private information and breached data protection rules.

Why "design" is the whole case

This is the part that makes the case matter beyond one MP, and it is easy to miss.

The usual defence of a general-purpose tool is that users decide what to do with it. Asato's claim attacks that directly.

"Grok is not a passive tool at the mercy of its users," she said. "It is engineered and designed to turn a two-word prompt into fake but realistic, degrading imagery."

6 of 21 posts still live
Reported in February, still up in late July. Photo: Tilman2007, Wikimedia Commons.

What the company admitted in its defence

According to the MP, SpaceXSI — formerly known as xAI — filed a defence on 30 October in which it admits the chatbot is designed to expand on short user prompts by pulling data from X threads and the wider internet to create deepfakes.

That admission is about the architecture, not about any individual user, which is precisely the ground Asato has chosen to fight on.

The sequence, as set out in court

Marie Demetriou KC, for Ms Asato, said the MP posted on X on 6 January calling on Musk to "understand that using AI to strip women without their consent is sexual assault designed to degrade and humiliate".

In the days that followed, several X users prompted Grok to create non-consensual images of Ms Asato, which were then circulated. Court documents published earlier this year said she was "victimised" after speaking out against the creation of such images of women.

The company also ran a "spicy mode" for Grok that produced more sexual content, with system prompts containing the instruction "no restrictions on adult sexual content or offensive content", according to the MP. In January the company changed that mode so it was available only to paying subscribers.

The moderation record

After multiple attempts to report posts, X refused to remove them, stating in March that they did not violate its terms of service, privacy policy or rules.

Demetriou said two videos were later removed, but six of 21 posts reported in February were still live in late July.

What is being asked for

In the UK it is illegal to create or request a non-consensual deepfake image of an adult. Asato is also asking the High Court to order SpaceXSI to "comply with existing laws to make its system safe for all".

"If an AI product cannot be deployed without automating the humiliation of women at scale, it should not be deployed at all," she said.

"I am taking this case forward to establish a clear principle: tech companies must be held accountable for the harm their systems actively generate."

Whether a court accepts that principle is the thing to watch. If design choices can carry liability, the reasoning would not stop at one chatbot.

Source: BBC

Written by

Tech BD

Tech BD is an independent technology publication based in Dhaka, published in English and Bangla. Articles appear under the masthead rather than under individual bylines; corrections and source questions reach the editor at info@tech-bd.com.