LONDON — Elon Musk’s AI chatbot, Grok, has implemented restrictions preventing non-subscribers from generating or modifying images. This decision follows widespread outrage over explicit deepfakes circulating on the platform, particularly involving women and, alarmingly, some of these images appeared to depict minors.
For the past few weeks, researchers noted a surge in malicious requests made by users, prompting various governments to condemn the platform and initiate investigations.
In response to the uproar, Grok announced on Friday that image generation and editing features are now exclusive to paying subscribers. Users are prompted to subscribe in order to access these functionalities.
While Grok’s specific subscriber numbers are unclear, there was a noticeable reduction in the generation of explicit deepfakes shortly after the announcement compared to days prior. However, the chatbot continues to process image requests, but only from users with blue checkmarks, typically granted to those who pay $8 monthly for premium features.
Despite these changes, European leaders and regulators remain unconvinced. “This doesn’t resolve our core concern. Whether users pay or not, we don’t want to see such images. It’s straightforward,” stated Thomas Regnier, a spokesperson for the European Union’s executive Commission. The Commission had previously criticized Grok for its troubling content.
The British government echoed similar sentiments, expressing dissatisfaction with Grok’s latest measures. Geraint Ellis, a spokesman for British Prime Minister Keir Starmer, described the changes as “not a solution,” suggesting they trivialize the experiences of victims of sexual violence. He asserted that X needs to act decisively and warned that all options are available if issues persist.
Starmer, speaking to a radio station, stressed the urgency for X to remove inappropriate content, stating, “It’s simply not acceptable.”
This week, U.K. media and privacy regulators reached out to both X and Musk’s AI firm, xAI, seeking details on compliance with British regulations, while countries like France, Malaysia, and India have also begun scrutinizing the platform. In Brazil, a lawmaker has called for an investigation. Meanwhile, the European Commission mandated that X retain all internal documents related to Grok until the end of 2026 as part of a broader inquiry under the EU’s digital safety regulations.
Grok is accessible to X users without charge, allowing them to pose questions either directly or in responses to other users. Launched in 2023, Grok included an image-generating capability labeled as “Grok Imagine,” which features a so-called “spicy mode” for producing adult content.
The situation is exacerbated by Musk marketing Grok as a bolder alternative to competitors with stricter rules. Additionally, the publicly visible nature of Grok’s images allows for easy dissemination, raising further concerns.
