Elon Musk’s Grok AI image editing restricted to paying users after deepfake

Elon Musk’s X platform, which has faced criticism for allowing people to create pornographic deepfakes, has limited the image-editing capabilities of its artificial intelligence tool Grok to paying subscribers.

The chatbot sparked a backlash after accepting users’ requests to take off other people’s clothes and digitally alter their images without their consent.

But Grok is now telling people asking it to produce such material that only paying subscribers can do so – meaning their names and payment information must be on file.

The BBC has contacted X for comment.

Those who don’t subscribe can still use Grok to edit images in its separate app and website.

Professor Claire McGlynn, an expert on the legal regulation of pornography, sexual violence and online abuse, said: “Musk threw his toys out of his pram in protest at being held accountable for the tsunami of abuse.”

“Instead of taking responsible steps to ensure that Grok was not used for abusive purposes, it removed access from the vast majority of users.”

This is after Govt. Regulator Ofcom is urged to use all its powers – up to and including an effective injunction – against X over concerns over illegal AI images created on the website.

Prime Minister Sir Keir Starmer said it was “disgraceful” and “disgusting” over concerns Grok had produced pornographic images of adults and children.

He said Ofcom had the “full support” of the government in taking action on content.

“It’s illegal. We won’t tolerate it. I’ve asked that all options be put on the table,” he told Greatest Hits Radio.

Government sources told BBC News: “We expect Ofcom will use all the powers at its disposal to deal with Grok and X.”

Ofcom’s powers Cybersecurity Law This includes the ability to seek court orders to prevent third parties from helping Elon Musk-owned platforms raise funds or be accessible in the UK.

The BBC has contacted the regulator for comment.

Grok is a free tool that allows users to request specific responses by tagging them directly in a post or in a reply under another user’s post.

But the feature also allows people to ask it to edit images and ask it to digitally remove most of people’s clothes.

Grok has fielded numerous requests from users to edit images of women to make them wear bikinis or tiny outfits – with those who have received such requests telling the BBC that it makes them feel “humiliated” and”dehumanizing“.

However, as of Friday morning, Grok has told users who ask it to make changes to images uploaded to X that “image generation and editing is currently limited to paying subscribers.”

It added that users “can subscribe to unlock these features.”

Some posts seen by BBC News on the platform suggest that only those with a blue checkmark “verified” mark (exclusive to X’s paid subscriber tier) can successfully request image edits from Grok.

Professor McGlynn said this move echoed X’s approach Last year, a Taylor Swift pornographic deepfake appeared on the platform – It blocks searches for pornographic content generated by pop stars using the Grok AI video feature.

“He did this to inflame free speech debates,” she added.

“He will claim that regulation is inhibiting people’s use of the technology. But all that regulation requires is that he take the necessary precautions to reduce harm.”