Elon Musk’s Grok chatbot has become a fresh stress test for Europe’s ability to regulate artificial intelligence, after the tool generated thousands of sexualised and digitally altered images of women and minors, prompting alarm among regulators and governments.
Late Wednesday, Musk’s AI company xAI said it had restricted image editing features on Grok following mounting pressure from authorities. The decision came after the chatbot was found producing hyper realistic images that portrayed women in degrading poses and minors digitally altered into revealing clothing. Musk had initially downplayed the issue, but the reversal underscored how difficult it has become to police AI systems that make explicit content fast, cheap, and widely accessible.
The controversy adds to a growing list of clashes between Musk and European regulators, spanning election integrity, content moderation, and the boundaries of free expression online. Many watchdogs say they are still struggling to keep pace with rapidly evolving AI tools, particularly when it comes to defining nudity, consent, and accountability between platforms and users.
Read also: Elon Musk Expands xAI Colossus As Power Use Nears 2 Gigawatts
British media regulator Ofcom welcomed xAI’s move but said its investigation into the matter was far from over. The agency said it was seeking detailed explanations on how the failures occurred and what safeguards would prevent a repeat. Earlier this month, Grok-generated images circulated widely on X, depicting women and girls altered into sexualised scenarios, some showing signs of physical abuse.
While xAI said it is now blocking the creation of images featuring people in skimpy clothing in regions where such content is illegal, it did not clarify which jurisdictions were covered. Some countries, including Malaysia and Indonesia, imposed temporary bans on Grok, while regulators in the UK, France, and Italy launched formal probes.
European lawmakers have called for tougher enforcement of the EU’s Digital Services Act, warning that voluntary fixes may not be enough. The European Commission said it was prepared to deploy its full regulatory powers if the changes proved ineffective.
Legal experts note that while Britain and several EU states criminalise the sharing of non-consensual intimate images, victims still face heavy emotional and legal burdens when seeking redress. Deepfakes existed long before AI chatbots, but Grok’s integration with X has given such content unprecedented reach.
British Prime Minister Keir Starmer welcomed the restrictions but warned that free speech does not extend to violating consent, adding that governments stand ready to strengthen laws if current measures fall short.