X has continued to allow users to post highly sexualised videos of women in bikinis generated by its AI tool Grok, despite the company’s claim to have cracked down on misuse.
The Guardian was able to create short videos of people stripping to bikinis from photographs of fully clothed, real women. It was also possible to post this adult content on to X’s public platform without any sign of it being moderated, meaning the clip could be viewed within seconds by anyone with an account.
It appeared to offer a straightforward workaround to restrictions announced by Elon Musk’s social network this week. These had been welcomed by the prime minister, Keir Starmer, who had described the photographs generated by Grok as “disgusting” and “shameful”.
After weeks of rising public concern, X said late on Wednesday it had “implemented technological measures to prevent the Grok account from allowing the editing of images of real people in revealing clothing such as bikinis”.
It said the restriction would apply to all users, including paid subscribers, and it had “zero tolerance for any forms of child sexual exploitation, nonconsensual nudity, and unwanted sexual content”. But it did not specify whether people would still be able to create such images on the standalone Grok app, and then share this material publicly on X.
The Guardian found that this standalone version of Grok, known as Grok Imagine – which is easily accessible through a web browser – was still responding to prompts to digitally remove the clothes from images of women.
Reporters uploaded still images of fully clothed real-life women, and prompted the AI tool to dress them in bikinis. The platform responded by going further than the request, to create short videos of the women removing their clothes in the manner of a sexually provocative striptease.
X has been contacted for comment.
Rebecca Hitchen, the head of policy and campaigns at the End Violence Against Women Coalition, said such a simple workaround should not be possible.
“The continued ease of access to sophisticated nudification tools clearly demonstrates that X isn’t taking the issue of online violence against women and girls seriously enough,” she said.
Hitchen called on the UK government and Ofcom, the media regulator, to pressure X and other platforms “to stop the proliferation of image-based sexual abuse”.
Reuters has also reported that its journalists, including a reporter in Britain, used Grok to create on-demand sexualised photos after the announcements by X.
“It’s hard to believe that xAI and Elon Musk can’t work out how to prevent these images from being spewed out by Grok,” said Penny East, the chief executive of the Fawcett Society. “First, Musk decided the solution was to preserve nudification as a privilege only for those users who pay for X. Then he pledged to stop it entirely. And yet it has not stopped.
“The truth is Musk and the tech sector simply do not prioritise safety or dignity in the products they create. It’s a pretty low bar for women to expect that they can converse online without men undressing them. And yet seemingly even that is impossible.”
While Downing Street had said it felt “vindicated” by the steps X had taken, there was also caution among ministers about the extent of the changes and how they would be delivered.
On Thursday, Liz Kendall, the technology secretary, who had described the sexual manipulation of images of women and children as “despicable and abhorrent”, welcomed the move and thanked “those who have spoken out against this abuse, above all the victims”.
However, she added: “I will expect the facts to be fully and robustly established by Ofcom’s ongoing investigation.”
Starmer also demanded X to act without delay. “Free speech is not the freedom to violate consent,” he said. “Young women’s images are not public property, and their safety is not up for debate. I welcome that X is now acting to ensure full compliance with UK law – it must happen immediately.”
Ofcom said its formal investigation into X, launched on Monday, remained ongoing and it was “working round the clock to progress this and get answers into what went wrong and what’s being done to fix it”.
Canada’s privacy watchdog said it was investigating xAI, while authorities in the Philippines said they were moving to block Grok, with Malaysian authorities planning to take legal action.
But the controversy may have been helpful for boosting public awareness of Grok. On Thursday, Musk shared a post claiming “popularity and real world usage are skyrocketing globally” – alongside a graph of “Grok” as a search term hitting a new high on Google Trends. Musk simply added: “Try Grok.com.”
A government spokesperson said: “The Online Safety Act already requires platforms like X to prevent illegal content, including nonconsensual intimate images and child abuse material, from appearing on their services.
“The secretary of state has said she expects X’s compliance with UK laws to be fully and robustly established by Ofcom’s investigation, which is already under way, and the government will not rest until social media companies meet their legal duties.
“We are also taking further action through the proposed ‘nudification’ offence, which will target tools designed specifically to generate nonconsensual intimate images.”
