Technology & Science
UK Pressure Forces X to Disable Grok’s “Digital Undressing” Feature
After Ofcom opened a probe and PM Keir Starmer threatened swift sanctions, X on 15 Jan 2026 geoblocked and technically disabled Grok’s ability to strip or sexualize images of real people, even for paying users.
Focusing Facts
- Ofcom’s formal investigation began 13 Jan 2026 and can levy fines up to 10 % of X’s global revenue or £18 million under the Online Safety Act.
- The new policy forbids Grok from editing any real person into bikinis/underwear worldwide and confines all remaining image-generation tools to verified premium subscribers.
- By mid-January 2026 Indonesia and Malaysia had already blocked Grok entirely, while California’s Attorney General and three U.S. senators launched inquiries of their own.
Context
Governments jerking a tech novelty back into compliance echoes Atlantic City’s 1893 ban on Kodak snapshot cameras that suddenly made beachgoers’ bodies public property, and the 2015 wave of U.S. “revenge-porn” statutes that shifted liability from victims to platforms. Starmer’s ultimatum highlights two century-long trajectories: every breakthrough in visual media first democratizes creation, then encounters a legal catch-up phase; and sovereignty over information has steadily migrated from companies’ terms of service to extraterritorial regulators armed with fines big enough to shape code. Whether today’s clamp-down endures will determine if generative AI follows the laissez-faire path of early social media or the tightly policed model of broadcast TV; in a 100-year frame, the precedent that algorithms—not just users—bear legal risk could close the “move fast, break things, ask forgiveness” era that defined the early 21st-century internet.
Perspectives
British mainstream and international general news outlets
e.g., Bloomberg Business, WION, The Irish News — They depict Keir Starmer’s threatened crackdown on Grok as a necessary, principled defense of women and children, welcoming X’s concessions as proof that strong state action works. Reliance on government quotes lets ministers take a victory lap, so stories accentuate political success and may gloss over free-speech or technical complexities.
Tech watchdog and consumer press
e.g., Metro, TechSpot, AppleInsider — Reporting stresses that Grok still produces sexualised deepfakes despite new rules, casting X’s fixes as inadequate and calling for even tougher measures or bans. By highlighting the most sensational failures they drive outrage-clicks, sometimes downplaying any real improvement and framing regulation as the only solution.
Tech industry trade press wary of regulation
e.g., Telecoms.com — They frame X’s tweaks as largely symbolic and mock the UK for ‘self-congratulation,’ arguing that broader global and market pressures—not politicised regulators—forced Musk’s hand. Scepticism toward regulators can tip into minimising victim harm, presenting safety rules as mere political theatre rather than a public-interest necessity.