The EU launches a formal probe into X over Grok-generated sexual deepfakes, testing how far platform liability extends under Europe’s AI and content safety rules.
 |
| A widening EU investigation into X signals a tougher line on generative AI, platform accountability, and the risks of sexual deepfakes. Image: CH |
Brussels, Belgium — January 27, 2026:
The European Commission’s decision to open a formal investigation into X over the alleged use of its AI chatbot Grok to generate sexually explicit deepfake images marks a decisive moment in Europe’s attempt to assert control over fast-moving generative AI technologies. While the case is rooted in specific allegations, its implications reach far beyond one platform, raising fundamental questions about responsibility, enforcement and the limits of innovation in the digital public sphere.
At issue is whether X breached its obligations under the EU’s Digital Services Act (DSA) by failing to prevent Grok from producing manipulated sexual images of real people and making them accessible to users within the bloc. If violations are confirmed, the company could face fines of up to 6 percent of its global annual revenue—a penalty designed to ensure that compliance costs are no longer treated as a manageable business expense by large platforms.
The investigation follows similar action by the UK communications regulator Ofcom and comes amid growing political and public alarm over AI-generated sexual deepfakes. Campaigners and victims argue that such content represents a uniquely damaging form of abuse, one that can be produced at scale, spreads rapidly and leaves those targeted with little practical recourse. European Commission Executive Vice-President for Technology Henna Virkkunen described the practice as harmful and degrading, emphasizing the need to protect users, particularly women and children.
X has said it blocked Grok from digitally altering images to remove clothing in jurisdictions where such content is illegal. EU regulators, however, appear unconvinced that these safeguards were either effective or consistently enforced. Irish Member of the European Parliament Regina Doherty said investigators would examine whether explicit manipulated images were accessible to users inside the EU, highlighting a recurring regulatory concern: that platforms announce guardrails without proving they work in practice.
The Commission has also warned it may impose interim measures if X fails to introduce meaningful protections quickly. This tougher stance reflects a broader shift in Brussels from guidance to enforcement, as well as frustration with what regulators see as reactive, rather than preventative, approaches to online harm. The probe has been widened to include risks linked to X’s content recommendation algorithms, reinforcing the view that AI tools and platform amplification cannot be regulated in isolation.
Elon Musk’s response has been characteristically combative. He has publicly mocked new restrictions related to Grok and accused regulators, including the UK government, of using safety concerns as a pretext for censorship. Those claims have found some sympathy among free speech advocates, but they clash with Europe’s regulatory consensus that platforms of X’s scale must shoulder proportionate responsibility for systemic risks.
The scale of Grok’s deployment adds urgency to the investigation. X has claimed the chatbot generated more than 5.5 billion images in a single 30-day period, illustrating how rapidly potential harms can spread if controls fail. Several countries, including Australia, France and Germany, are examining the chatbot’s operations, while Indonesia and Malaysia previously imposed temporary bans—signs that the unease is global rather than uniquely European.
The probe also comes just weeks after the EU fined X €120 million over concerns that its blue-tick verification system misled users. U.S. officials have criticised the EU’s actions, accusing European regulators of unfairly targeting American technology firms—an argument echoed by Musk. Brussels rejects that narrative, insisting its enforcement is based on risk and market impact, not nationality.
Ultimately, the Grok investigation is less about a single chatbot than about setting precedents. As generative AI becomes embedded across social platforms, the EU is signaling that innovation does not dilute legal responsibility. Whether X adapts or continues to challenge regulators, the outcome is likely to shape how AI-powered services operate—not just in Europe, but worldwide.