The AI technology has been used to undress people without their permission
Concerns over generative AI took a darker turn this week as Grok, the AI chatbot developed by Elon Musk’s xAI and integrated into X, was accused of continuing to generate non-consensual sexual deepfakes of women — and, increasingly, children — despite months of warnings from researchers and regulators.
The controversy comes amid reports that xAI has raised a further $20bn in funding, a claim that has intensified criticism from digital safety advocates who argue that financial success is being prioritised over basic safeguards.
Pictured top: Some of the Grok edits found online Left: Dot Branning’s head from EastEnders, but an AI generated body. Right: Meg Richardson of Crossroads gets a fake bikini body.
Jurgita Lapienytė, editor-in-chief of cybersecurity publication Cybernews, warned that the scale and accessibility of the technology marks a dangerous shift. “Every woman with a photo online is now a target,” she said, unless the system is disabled or radically altered. “Grok has taken something that once required technical skill or fringe platforms and placed it directly onto one of the world’s largest social networks.”
At the centre of the issue is Grok’s image-generation and editing functionality, which critics say has been repeatedly exploited to bypass safeguards and produce explicit, non-consensual images. Victims include journalists, politicians and activists, with several cases involving minors now under investigation.
While xAI has described the incidents as safeguard failures, Lapienytė argues the problem can no longer be framed as accidental. “This didn’t start yesterday,” she said. “When The Verge revealed Grok generating unsolicited nude videos of Taylor Swift back in August, that should have been the moment to stop. Watchdog groups warned exactly what would happen if it wasn’t fixed. Five months later, it’s worse.”
She points to the growing number of cases involving children as young as 10 as evidence that the risks were both foreseeable and preventable. “At some point, we have to stop calling this a failure and start calling it a deliberate choice to ignore warnings,” she said.
Campaigners are particularly alarmed by the speed at which the technology has been normalised. “If regulators don’t force X to disable this immediately,” Lapienytė warned, “we’re looking at the normalisation of on-demand sexual exploitation at a scale we’ve never seen before.”
The fear is that Grok’s integration into a mainstream platform lowers the barrier for abuse. Where deepfake pornography once required technical expertise or paid services operating in legal grey areas, critics say Grok allows users to generate and distribute such content within minutes — and in public view.
“If X gets away with this,” Lapienytė added, “every other AI company will see there’s no consequence for building similar tools. Within months, we could see this weaponised for revenge porn, sextortion, harassment of minors and blackmail.”
The backlash has prompted scrutiny from regulators in the UK, EU, India and Australia, all of whom have demanded explanations from xAI and X over how safeguards failed — and why they were not strengthened sooner. Ofcom is reportedly assessing whether the platform has breached UK online safety obligations.
In a rare and striking moment of self-critique, Grok itself acknowledged the harm caused when responding to a prompt about one high-profile victim, journalist Evie Smith, who was targeted by trolls using a so-called “see-through bikini” loophole. The system described the incident as a “violation” and “coordinated abuse”, noting that such attacks disproportionately affect women who speak publicly about feminism or politics.
While xAI has since tightened restrictions, critics say the response came too late — after thousands of images had already circulated. “The technology doesn’t go back in the bottle,” Lapienytė said. “But we can still decide whether it lives on mainstream platforms or is pushed back into the dark corners of the internet where it belongs.”
For survivors and digital safety advocates, the question now is whether regulators will act decisively — or whether Grok will become a precedent for what happens when innovation races ahead of accountability.
“This isn’t about banning AI,” Lapienytė said. “It’s about drawing a line. Transparency without enforcement is meaningless, and safeguarding without consequences is just branding.”
As generative AI continues to reshape online spaces, Grok’s trajectory may prove a defining test of whether public platforms are willing — or able — to prioritise human harm over growth.
