The AI chatbot Grok, developed by xAI, became the center of an international scandal after enabling a service that generated sexual deepfakes (digitally “undressed” images of people) on the X platform. Experts from Check Point are sounding the alarm, warning that the vulnerability must be closed at the model level so that prompts and edits leading to the artificial “undressing” of users are blocked.
According to Check Point’s experts, platforms should treat this issue as a security incident and remove the flaw within the model itself to eliminate the possibility of distorting reality into sexualized content. They also point out that the lack of international treaties and harmonized AI regulations allows platforms such as X to try to avoid responsibility. “Partial restrictions will not protect victims whose images have already been created, shared, and stored,” emphasizes Charlotte Wilson of Check Point.
A similar stance has been taken by the Polish government. “Uncontrolled AI is running amok and causing increasing harm. Grok, once again—at Elon Musk’s inspiration—is testing how far it can go…. There is no consent for breaking the law with impunity and for AI to create illegal content that destroys people and their mental health,” said Poland’s Minister for Digital Affairs Krzysztof Gawkowski in a post on X.
“Let’s be clear: a system that insults, promotes sexual content, and violates human dignity is the result of poor governance and a lack of ethical standards on the platform,” the minister added.
The scale of the problem has prompted strong reactions from authorities worldwide. The European Union described Grok’s behavior as “illegal” and “repulsive.” Officials in France, India, and Malaysia have also spoken about investigations.
Grok was launched in 2023 as a “sharper” alternative to competing chatbots that feature stronger safeguards. Last summer, an image generator—Grok Imagine—was added, equipped with a so-called “spicy mode” that enabled the creation of adult content. This feature, combined with the public visibility of generated images, allowed controversial content to spread rapidly online. In response to the backlash and criticism, Grok began restricting image generation and editing for most users (currently available only to paying subscribers).
However, experts noted that in recent weeks Grok complied with numerous potentially malicious user requests—from “dressing” women in bikinis to placing them in explicitly sexual poses. In several cases, serious concerns were raised that the generated images could involve children.
Grok’s future in question
The scandal surrounding Grok shows how thin the line is between innovation and responsibility. Public visibility of content, a “bold” approach to so-called content independence, and insufficient safeguards have turned Musk’s chatbot into a symbol of the challenges facing today’s AI industry. The question is no longer whether regulation is needed—but how quickly and how effectively it will be implemented before technology causes further harm.
“Grok under pressure on many fronts: the EU, the United Kingdom, France, India—sanctions risk for X is growing” was first published in CEO Magazine.
Source: ceo.com.pl
