The social media network X has sparked significant anger after users used its AI chatbot Grok to manipulate pictures of women by removing their attire.
Multiple instances have emerged where the system has been used to virtually strip women to show them in bikinis without consent, as well as to put them in sexualized situations.
Developer Response and Legal Position
XAI, the firm responsible for Grok, did not respond to inquiries, except for an automatically-generated message stating "legacy media lies".
A Home Office representative stated that legislation is being proposed to ban image stripping software. Under a upcoming criminal offence, anyone who supplied such technology would "receive a prison sentence and substantial fines".
"X or Grok has the ability to stop these types of abuse if they wanted to," said a law professor, adding they "seem to have impunity."
Regulator Position and Tool Capabilities
The communications watchdog Ofcom said tech firms must "consider the danger" of people viewing unlawful material on their networks, but stopped short of confirming whether it was at this time investigating X or Grok in relation to AI-generated pictures.
Grok is a free chatbot—with some paid-for functions—which reacts to X users' queries when they mention it in a message.
It is commonly used to provide reactions or further context to other users' comments, but people on X are also capable of alter an posted photo through its AI image editing feature.
Background of Criticism and Policy
The chatbot has been faulted for allowing users to generate images and clips with nudity and sexualised content, and it was earlier alleged to have been involved in creating inappropriate material of celebrities.
"The platform has been permitting the generation and spread of these images for months without taking any action and we have not seen any intervention by oversight bodies," Professor McGlynn added.
XAI's own terms of service prohibits "depicting likenesses of people in a pornographic way".
In a comment, Ofcom said it was unlawful to "create or distribute non-consensual intimate images or child sexual abuse material" and stated this applied to sexual deepfakes made with AI.
It said platforms such as X were required to take "suitable actions" to "lessen the risk" of users coming across unlawful material on their platforms and to take down it quickly when they learn of it.
