Government accused of dragging its heels on deepfake law over Grok AI

Campaigners have accused the government of dragging its feet on legislation that would make it illegal to create sexualized deepfakes without consent.

It comes amid a backlash over images created using Elon Musk's AI Grok to digitally remove clothes, with one woman telling the BBC that more than 100 sexualized images were created of her.

Currently It is illegal to publish deepfakes of adults in the UKbut new legislation making it a criminal offense to create or request them has not yet come into force, despite will take place in June 2025..

But it is unclear whether all images of undress created by Grok would violate this law. The BBC has contacted the government for comment.

In the statementX said: “Anyone who uses Grok or encourages it to create illegal content will suffer the same consequences as if they had uploaded illegal content.”

It comes after Prime Minister Sir Keir Starmer, addressing concerns about Grok's sexualized images of adults and children, called it “disgraceful” and “disgusting”.

“This cannot be tolerated,” he told Greatest Hits Radio.

“X needs to get this under control, Ofcom has our full support to take action on this.

“This is wrong.”

Grok can be accessed through the website and app, and by tagging “@grok” in posts on the X social media platform.

But Andrea Simon, from End Violence Against Women, said the government had “put women and girls at risk” by failing to implement the law.

“Sexually explicit deepfakes made without consent are a clear violation of women's rights and have a long-term traumatic impact on victims,” she said.

“For women using platforms like X, the threat of such violence may also mean that they feel the need to self-censor and change their behavior, limiting their freedom of expression and participation online.”

“This is not only a criminal justice issue, but also a matter of regulating the technology ecosystem that facilitates and profits from violence against women and girls.”

On TuesdayTechnology Minister Liz Kendall demanded that X “get to the bottom of this urgently”, calling the situation “absolutely appalling”.

Regulator Ofcom said it had made “urgent contact” with X and xAI, which created Grok. on Monday and told the BBC it was investigating the concerns.

Both Kendall and Downing Street supported the regulator's action, with the Prime Minister's spokesman adding on Wednesday that “all options remain on the table”.

The Ministry of Justice told the BBC: “It is already an offense to share intimate images on social media, including deepfakes, without consent.

“We refuse to tolerate such demeaning and harmful behavior, which is why we have also passed legislation prohibiting their non-consensual creation.”

Current law states that it is illegal to create pornographic deepfakes when they are used in revenge porn or depict children.

A provision of the Data (Use and Access) Act 2025 criminalizes the creation or use of “purported intimate images” – or deepfakes, Lorna Woods, professor of internet law at the University of Essex, told the BBC.

But a year after the first announcement of the suppression of the crimeThe government has yet to introduce a key legal provision that could target those asking Grok to make sexualized deepfakes of women, experts and campaigners have told the BBC.

Professor Woods said the offense “appears to work well for some images created with Grok” but has not yet come into force.

Meanwhile, EVAW's Simon questioned why secondary legislation that would have criminalized the creation of sexual deepfakes had not yet been introduced, given that they were a “clear violation of women's rights”.

“This law has not yet come into force,” she said, “and the date when this will happen has not been determined.”

Conservative peer Baroness Owen, who campaigned for legal changes in the House of Lordsreported to the BBC that the government had “repeatedly delayed” bringing the rules into force.

“We cannot afford any further delays,” she said.

“The victims of this violence deserve better.

“No one should have to live in fear that their consent will be violated in such a horrific way.”

And backbencher Baroness Beban Kidron told the BBC: “Technology is moving fast and this legislation is meant to fill an existing gap, so there is no excuse for delay.”

BBC has talked to several women who had photos they posted on X that Grok's AI turned into deepfakes.

Grok is a free tool that users typically tag under other people's comments with hints about reactions or context.

More recently, people have been able to use a chatbot to edit photos they or others have uploaded, tagging them and adding text prompts, even if it involves stripping or posing in sexual positions.

Evie told the BBC that while she noticed that a few months ago people started replying to her photos asking Grok to put her in a bikini, the latest update made it “easier” for users and the photos looked “more realistic”.

She says she now has at least 100 sexualized images generated by the AI ​​assistant, so many that she refused to report them all due to the “mental strain” of viewing them.

In addition to having to look at the images herself, the thought of her loved ones seeing her in sexual or unclothed positions also makes it very difficult for her to be on X.

“My family is watching me, my friends, my colleagues,” she said.

“Knowing that all the people I care about in my life can see me like this… it’s disgusting.”

Dr Daisy Dixon, another X user, told the BBC she had noticed an increase in people using Grock to undress her, especially using her profile photo.

She said the photos left her feeling “humiliated” and said the fact that Grok automatically commented to her on altered images left her feeling like she was being attacked.

“Sending it back to you is like saying, 'I have control over you and I will continue to remind you that I have control over you,'” she said.

“We don't want to dilute the concept, but it feels like an attack on the body.”

And X users like Evie say action needs to be taken now.

“There are so many places on the internet where you can do this, but the fact that this was happening on Twitter using a built-in AI bot is crazy, it's allowed,” she said.

“Why is this allowed and why is nothing done about it?”