A recent attempt to address concerns about explicit imagery generated by the AI chatbot Grok appears incomplete, leaving significant vulnerabilities exposed. The initial statement focused on restrictions for users directly tagging the Grok account on the X platform, but conspicuously omitted any mention of the standalone Grok website and app.
While the company announced it would block the creation of problematic images within “Grok in X” – the chatbot integrated directly into the X platform – this block isn’t absolute. Instead, the imagery will be “geoblocked,” meaning access will be restricted only in regions where such content is legally prohibited. This approach raises questions about the protection offered to users in areas with less stringent regulations.
The situation is further complicated by apparent inconsistencies in the messaging. The announcement stated the Grok account wouldn’t accept certain requests, yet also mentioned geoblocking similar requests made through tagging. This feels redundant, given the initial claim of outright rejection.
Critiques haven’t centered on the generation of fully nude images, but rather on the creation of non-consensual explicit imagery, which is illegal in many places, including the UK, regardless of the level of nudity. This nuance is crucial, as the response appears narrowly focused.
Despite the announced crackdown, testing reveals ongoing issues. Journalists were still able to generate revealing deepfakes using the Grok app, which wasn’t addressed in the initial update. This suggests a significant gap in the implemented safeguards.
Personal experimentation confirmed these findings. Using a photograph as input, both the Grok app and the standalone website produced full-body deepfakes featuring revealing clothing not present in the original image. The in-X chatbot also altered poses to be more provocative, even without explicit prompting.
The battle to control the output of this AI is far from over. It remains unclear whether the omission of the app and website was an oversight, or a deliberate decision to address only the most visible points of concern.
The company’s stated commitment to “zero tolerance” for child sexual exploitation, non-consensual nudity, and unwanted sexual content feels undermined by these ongoing vulnerabilities. The current measures appear insufficient to uphold that pledge.
Requests for clarification from the company have been met with dismissive responses. When NBC News inquired, they were simply told “Legacy Media Lies,” raising concerns about transparency and accountability.
Meanwhile, pressure is mounting from lawmakers. U.S. Senators have sent a letter to Apple and Google, urging them to remove the Grok apps from their stores, citing clear violations of existing platform policies. This calls for immediate action beyond the company’s current efforts.
Governments are launching investigations, but the senators argue that waiting for official rulings isn’t enough. They believe the app stores have a responsibility to protect users now, by removing a platform demonstrably failing to adhere to established standards.
