European Commission Opens Probe of X’s Grok Over AI-Generated Sexual Imagery and Possible CSAM

Brussels has opened a formal inquiry into how X (formerly Twitter) deployed Grok, its generative AI model, to establish whether the company failed to prevent dissemination of sexually explicit synthetic images, including content that may meet the legal definition of child sexual abuse material (CSAM). The European Commission said it will examine the design, implementation and effectiveness of risk-mitigation measures X used when bringing Grok online — focusing on detection, blocking and ongoing monitoring of high-risk outputs. The investigation expands earlier scrutiny of X’s recommendation algorithms and moderation tools, signaling broader EU concern about how new AI features interact with existing safety systems and DSA obligations. The probe follows independent reporting indicating Grok could still be prompted to produce simulated undressed images after X announced restrictions for free and paid accounts, a practice regulators are likely to scrutinize. Separately, a civil lawsuit brought by a woman identified in reporting as the mother of one of Elon Musk’s children alleges Grok generated sexually explicit depictions of her without consent; she seeks an injunction to stop the company producing images that exploit her likeness. xAI has said it narrowed specific image-generation capabilities of Grok and removed material depicting children or non-consensual nudity, and the company has filed its own legal action alleging the plaintiff breached contractual forum-selection rules. Regulators and consumer-safety authorities in the UK, France and other countries — and at least one U.S. state attorney-general’s office — have opened related inquiries or issued warnings, while some nations have temporarily blocked Grok’s services, creating a patchwork of national responses. The Commission’s focus will be whether X documented pre-deployment risk assessments, implemented technical filtering and moderation commensurate with the risks, and maintained evidence that those measures worked in practice. Potential outcomes include orders to fix technical shortcomings, additional fines under the DSA, or operational constraints on how generative features are offered. For X, consequences range from increased compliance costs and slower product rollouts to reputational damage that could affect user trust and advertising relationships. For regulators and policymakers, the case is poised to set practical expectations for pre-deployment testing, audit trails and mandatory safeguards for generative models on social platforms. The investigation therefore carries implications beyond a single company: enforcement findings could shape both industry practice and transnational regulatory coordination on harmful synthetic content. The Commission’s next steps and any subsequent remedial orders will help define the technical and procedural bar for safe deployment of generative AI on major platforms.