AI & ComputeLOT 261120 · 3 min read

SpaceXSI Tells UK Court Obviously Fake Grok Images Cannot Breach Privacy

SpaceXSI denied liability for fake sexualised Grok images of Labour MP Jess Asato, telling the High Court that obviously fictitious AI images cannot breach privacy — a first UK test of AI developer liability.

By Grace Kim533 words

Key takeaways

  1. SpaceXSI filed its defence at London's High Court on Oct. 29; Reuters reported it on Nov. 19.
  2. Labour MP Jess Asato filed her claim in June, seeking damages and an injunction against Grok image generation.
  3. The company removed one video showing Asato "being chloroformed and prepared for a sexual assault" but says remaining images are obviously fictitious.
  4. xAI restricted Grok image editing in mid-January; a Reuters investigation in early February found sexualised images still possible.
  5. The case is among the first to test whether an AI developer is liable for system-generated content.
SpaceX's AI unit denies liability for Grok images of UK lawmaker, arguing obviously fake images cannot breach privacy
SpaceX's AI unit denies liability for Grok images of UK lawmaker, arguing obviously fake images cannot breach privacy — QuinceCreative / pixabay

SpaceX Super Intelligence, the Elon Musk company behind the Grok chatbot, has denied legal responsibility for fake sexualised images of British lawmaker Jess Asato, arguing that images any reasonable person would recognise as fictitious cannot violate privacy, Reuters reported on Nov. 19.

The company set out its position in a written defence filed at London's High Court on Oct. 29. Asato, a member of the governing Labour Party, sued in June. She wants an order to stop Grok creating manipulated images of her, plus damages and a declaration that the company acted unlawfully.

What does the company argue?

SpaceXSI denies Asato's claims of misuse of private information and breach of data protection law. It also contests her request for an order requiring it to "implement effective and permanent technical measures" to stop Grok generating manipulated images of her.

Its lawyers made three main arguments:

  • The company already removed AI images of Asato that broke its terms, including a video showing her "being chloroformed and prepared for a sexual assault".
  • Other images left online "would be understood by reasonable and sensible persons viewing that content to be AI-generated, fictitious and implausible ... rather than as realistic images", so Asato cannot claim a reasonable expectation of privacy.
  • On data protection, SpaceXSI "did not determine the content of the relevant users' prompts".

The privacy argument is the sharpest point in the filing. If a court accepts that visibly implausible AI images fall outside privacy law, victims of non-consensual sexualised imagery would face a much harder path to injunctions against image generators.

How did Asato respond?

Asato rejected the company's framing on Thursday. "SpaceX SI wants to have it both ways," she said. "The company admits it built the AI model, rewrote user requests to add its own graphic details, pulled images from across the internet, and published the results from an account it controls. Yet it also claims it is a neutral bystander without legal responsibility."

Her original claim, filed in June, alleges Grok was designed and trained in a way that let users generate sexualised images of her. The case turns that design question into a legal one.

Why does the case matter for AI developers?

The lawsuit asks a question UK courts have not yet answered: whether the maker of an image generator is liable for what its users produce. A ruling against SpaceXSI could shape how AI developers design image tools and handle takedown requests in the UK.

Grok is available through Musk's social network X. It has drawn regulatory scrutiny in several countries over non-consensual sexualised images. xAI said in mid-January it had restricted image editing and blocked images of people in revealing clothing "where it's illegal". A Reuters investigation in early February found Grok still produced such images after those restrictions.

Musk's AI group has faced other recent setbacks. They include a Grok Build security lapse and job cuts on the Grok Imagine video team.

The High Court has not set a date for the next hearing. Whatever it decides will set the first UK precedent on whether an AI developer answers legally for the images its model produces.

By Grace Kim

mlq.ai (Source)

Filed under

Share this article:

More from Grace Kim

Grace Kim

Show full bio

Market editor covering people, launches and day-to-day practice in Semiconductors, chip manufacturing, hardware, AI compute at Chip Dispatch.

1,170 articles

Read also

« Previous article