X restricts Grok AI from generating bikini images of real people amid regulatory pressure
X, formerly Twitter, has restricted its Grok AI tool from generating or editing images of real people in bikinis, underwear, or similar revealing clothing in jurisdictions where such content is illegal. The move follows mounting criticism over sexualised AI deepfakes and regulatory pressure in the US and UK. The announcement came hours after California authorities disclosed an investigation into such AI-generated images, including those involving children. The UK government welcomed the decision, with Prime Minister Sir Keir Starmer having publicly urged X to rein in its tools. The UK's Ofcom regulator, which is investigating whether X breached UK law, called the change a 'welcome development' but stressed that its probe continues. X stated that technological measures and location-based controls have been implemented to geoblock these capabilities. The company also noted that only paid users can access image editing features on Grok, aiming to improve accountability. Elon Musk defended the platform's NSFW settings, stating they permit limited, fictional content only, but drew backlash for posting AI-generated images of PM Starmer in a bikini. If found non-compliant, Ofcom could seek a court order to block X in the UK. The issue erupted in late 2025 when users created non-consensual sexualised images of real people, raising alarms about child sexual abuse material. In response, X had previously restricted Grok, but investigations continue. In the US, California’s attorney general is probing Grok’s deepfakes for violations. Other nations, including Malaysia and Indonesia, have temporarily blocked Grok entirely over concerns about sexually explicit AI content harming citizens, especially minors.