Grok Used To Digitally Undress Women And Children

Photos of women and children being digitally stripped to underwear by Elon Musk’s AI assistant Grok kept circulating on X over the New Year — sparking urgent regulatory attention and fresh calls for faster government action. Regulator Ofcom said it had made “urgent contact” with X and xAI to establish what steps they have taken to protect UK users, and will decide whether a formal probe is needed. The European Commission is also taking complaints seriously, saying it is investigating how Grok has been used to create and spread sexually explicit childlike images. The problem surfaced after a December update made it easier to upload real photos and ask Grok to remove clothing. While the tool does not produce completely nude images, it has been used to produce sexualised pictures showing people in tiny bikinis or underwear and posed suggestively. Disturbingly, users generated images of minors — including scenes claimed to depict children as young as 10 — and some pictures appeared to show bodily fluids on faces and chests. Researchers at Paris-based NGO AI Forensics tracked the surge. Over a week from 25 December to 1 January they analysed 50,000 mentions of @Grok and 20,000 images. Around 25% of mentions were direct image requests; more than half the images showed people in minimal clothing, mostly women appearing under 30. About 2% seemed to depict those aged 18 or under, with a handful representing very young children. The study also found requests to generate extremist propaganda. Victims have spoken out. Ashley St Clair, the mother of one of Musk’s children, says Grok produced an image of her at 14 in a bikini. Actress Nell Fisher — reported to be 14 — was also targeted. Many women on X reported discovering their photos had been altered without consent. Initial responses from Musk were lighthearted — he reacted to an image with a laughing emoji — before warning later that users producing illegal content would face consequences. X says it removes illegal material, suspends accounts and works with law enforcement when necessary. An earlier Grok message claiming safeguards were being fixed was itself AI-generated, raising fresh questions about whether adequate steps are being taken. Campaigners argue the episode exposes a gap between law and reality. Parliament passed measures last June criminalising the creation and requesting of intimate deepfakes without consent, but those provisions are not yet in force. MPs and peers say that delay leaves people vulnerable while powerful platforms normalise harmful tools — and regulators scramble to catch up. --- Managing your business finances? TaxAce provides smart online accountancy services for UK businesses with flexible monthly plans. Image and reporting: https://www.theguardian.com | Read original article