It urged ministers to pass new legislation explicitly covering generative AI services – a recommendation the government rejected a month later, on the basis that amending the Act before it was fully in force risked ”complicat[ing] and undermin[ing]” implementation.
“The vibe has very much been one of ‘trust the process,’” Owen Bennett, who led international online safety at Ofcom until December 2025 and now works as an independent digital policy consultant, said. Various parts of the OSA took time to come into force, but now key duties are in force, that “grace period” is passed, Bennett said.
Pre-Grok, the U.K. government was also clearly signaling that it didn’t intend to follow Australia’s lead in banning social media for children, with a spokesperson for No.10 Downing Street saying in December there were no plans to implement a ban and “it’s important we protect children while letting them benefit safely from the digital world, without cutting off essential services or isolating the most vulnerable.”
Victory lap
Although Musk was initially defiant, X did eventually back down. Faced with shutdowns and threats of legal action, X agreed to restrict Grok’s image-generation function in jurisdictions where sexualized deepfakes are illegal.
Starmer claimed a victory, and he hasn’t stopped claiming it. His hardline approach to Grok appeared to enjoy public backing: polling by More in Common in January, reported by The Guardian, found that 58 percent of Brits thought X should be blocked in the U.K. if it didn’t crack down on AI-generated nonconsensual images.
The fight with Grok and X has become a frequent refrain. As recently as last month Starmer referred to “the fight that we had with Grok” in the Commons. Just last week, Tech Secretary Liz Kendall listed “standing up to Grok and X” as an example of how she is protecting children online.
