This Isn’t Spicy – It’s Illegal: The EU Condemns Musk’s AI Grok

The internet has always had a dark side, whether you’ve seen it or not. But every so often, a new technology forces all of us to confront just how quickly things can go wrong when reasonable safeguards fail, and when permissiveness is treated as a feature rather than a risk. That’s exactly what’s happening now with Grok, Elon Musk’s AI chatbot, as European regulators move quickly to investigate its role in generating inappropriate, sexualized images of unknowing women and children.

What’s Happening?

In recent weeks, Grok (a free AI assistant created by Elon Musk) has come under scrutiny after users on X, formerly Twitter, began exploiting a newly released feature that allows users to edit the pictures they upload. What was introduced as a good-faith creative tool turned into something far more disturbing. People discovered that they could ask Grok to digitally remove clothing and other barriers from photos (like emojis, which parents often use to cover the faces or private parts of their kids online), most often involving women, and in some reported cases, minors, producing sexualized, non-consensual images that then get circulated online.

This isn’t an isolated incident that’s been blown out of proportion – Reuters reviewed public prompts sent to Grok during a single 10-minute window and counted over a hundred attempts to generate images showing people as at least partially undressed. That volume alone should terrify everyone, because it suggests an exploitable systemic failure, not a rare edge case.

The World is Fighting Back

The European Union is taking this situation very seriously. EU officials have publicly condemned Grok’s so-called “Spicy Mode”, stating that when sexualized content includes imagery, including kids, the issue isn’t concerned with morality – it’s straight up illegal. Under the Digital Services Act (DSA), large online platforms like X are required to mitigate the risk of unlawful content spreading, including child sexual abuse (CSA) material and non-consensual sexual imagery. From the EU’s perspective, Grok appears to have neglected its responsibility.

Individual countries are following suit. France has reported X to prosecutors, while the UK has made urgent contact with X and xAI to investigate potential breaches of the Online Safety Act. Beyond Europe, India and Malaysia have issued takedown orders and launched their own investigations. What began as a platform-level scandal has now exploded into a global regulatory phenomenon. And it’s time to act now.

Let me be absolutely clear: neither I nor any reasonable person supports this kind of content. Sexualizing people without their explicit consent is abuse. Sexualizing children, whether they’re real or AI-modified, is horrifying, indefensible, and illegal. There’s no free speech loophole here, and there’s no technological excuse.

Rule 34 and Internet Culture

Part of this problem is cultural. Internet history is steeped in what’s often called “Rule 34”. It’s the cynical idea that if something exists, there will be pornographic content of it online, whether that’s children’s cartoon characters (like the infamous My Little Pony images), celebrities (everyone from Millie Bobby Brown to Emma Watson), or inanimate objects like aliens or Mr. Peanut. What started as an observation of people’s behavior has become an inevitability, an excuse, a justification. This, when combined with generative AI tools, is not only predictable but also dangerous. Normalizing abuse only ensures that it keeps happening. 

What makes this scandal especially troubling is that it was avoidable. Grok has been positioned as a more permissive AI when compared to mainstream AI models, which impose stricter limits on sexual and image-based content. While drawing lines between adult nudity, suggestive material, and illegality is complex, content involving minors and non-consensual sexual imagery isn’t a gray area. If an AI system can’t reliably enforce those boundaries, it shouldn’t be deployed at scale.

Why This Matters

This controversy highlights a truth about the internet: tools are neutral, but how they’re built and handled determines impact. Technologies enabling creativity and expression can just as easily be misused for harassment and exploitation, especially against women and children, who are already disproportionately targeted online.

Condemning abuse doesn’t mean supporting blanket restrictions. Shutting tools down or limiting access in specific regions wouldn't solve the problem; it would just move it elsewhere. It’s good that problems are being called out, but real progress will depend on platforms taking responsibility and fixing the issue at the source. If safety remains an afterthought rather than a core principle, the abuse will continue regardless of who’s blocked from accessing the tool.