Elon Musk’s platform X has limited image editing with its AI tool Grok to paying users, after it came under fire for allowing people to make sexualised deepfakes.
There has been a significant backlash after the chatbot honoured requests from users to digitally alter images of other people by undressing them without their consent.
But Grok is now telling people asking it to make such material that only paid subscribers would be able to do so – meaning their name and payment information must be on file.
The BBC has approached X for comment.
Those who do not subscribe can still use Grok to edit images on its separate app and website.
“Musk has thrown his toys out of the pram in protest at being held to account for the tsunami of abuse,” said Professor Clare McGlynn, an expert in the legal regulation of pornography, sexual violence and online abuse.
“Instead of taking the responsible steps to ensure Grok could not be used for abusive purposes, it has withdrawn access for the vast majority of users.”
It comes after the government urged regulator Ofcom to use all its powers – up to and including an effective ban – against X over concerns about unlawful AI images created on the site.
Addressing concerns that sexualised images of adults and children had been generated by Grok, Prime Minister Sir Keir Starmer said it was “disgraceful” and “disgusting”.
He said Ofcom had the government’s “full support” to act on the content.
“It’s unlawful. We’re not going to tolerate it. I’ve asked for all options to be on the table,” he said in an interview with Greatest Hits Radio.
Government sources told BBC News: “We would expect Ofcom to use all powers at its disposal in regard to Grok and X.”
Ofcom’s powers under the Online Safety Act include being able to seek a court order to prevent third parties from helping the Elon Musk-owned platform raise money or be accessed in the UK.
The BBC has approached the regulator for comment.
Grok is a free tool which users can tag directly in posts or replies under other users’ posts to ask it for a particular response.
But the feature has also allowed people to request it to edit images – and ask it to digitally strip people of most of their clothing.
Grok has fulfilled many user requests asking it to edit images of women to show them in bikinis or little clothing – something those subject to such requests have told the BBC left them feeling “humiliated” and “dehumanised“.
However as of Friday morning, Grok has told users asking it to alter images uploaded to X that “image generation and editing are currently limited to paying subscribers”.
It adds users “can subscribe to unlock these features”.
Some posts on the platform seen by BBC News suggest only those with a blue tick “verified” mark – exclusive to X’s paid subscriber tier – were able to successfully request image edits to Grok.
Prof McGlynn said the move echoed X’s approach to pornographic Taylor Swift deepfakes on the platform last year – where it blocked searches for sexualised material generated of the popstar using a Grok AI video feature.
“He is doing this to stoke free speech arguments,” she added.
“He will claim regulation is stifling people’s use of this technology. But, all the regulation requires is that he takes necessary precautions to reduce harm.”
