EU lawmakers just voted to ban nudify apps — and Musk’s Grok is Exhibit A

The European Union is moving fast to close a glaring loophole in its AI Act, and Elon Musk’s Grok chatbot just became the poster child for why that loophole exists in the first place.

This week, the European Parliament’s Internal Market and Civil Liberties committees voted 101–9 (with 8 abstentions) to simplify the AI Act and, more importantly, “propose bans on AI ‘nudifier’ systems.” The joint press release makes clear that the target is software that generates sexually explicit deepfake nudes — including images of real people and, horrifically, children.

The vote didn’t come out of nowhere. Earlier this year, the European Commission quietly admitted that the AI Act as written does not prohibit “AI systems that generate child sexual abuse material (CSAM) or sexually explicit deepfake nudes.” That’s a stunning gap for a law that was supposed to be the world’s most comprehensive AI regulation. Parliament members had already started drafting amendments, and this week’s vote formalizes their intent.

So why does Grok matter here? Because Musk’s chatbot became a prime example of exactly what happens when an AI platform fails to block outputs that sexualize real people. Users found ways to prompt Grok into generating non-consensual intimate images — including of minors. Musk’s response? Blame the users. That’s a familiar tactic in tech circles: deflect responsibility onto the people who abused the system, rather than admitting the system was designed without adequate safeguards.

But the EU isn’t buying it. The lawmakers behind this vote are explicitly citing Grok as evidence that voluntary guardrails don’t work. When a major platform with billions in resources can’t — or won’t — prevent its model from being weaponized for sexual abuse, the argument for mandatory, enforceable bans writes itself.

What’s interesting here is how quickly the EU is moving. The AI Act only entered into force last year, and already Parliament is amending it to address specific harms. That’s unusually fast for European bureaucracy, which suggests this issue has genuine political urgency. The 101–9 vote margin is about as close to unanimous as you get in a body this large.

Some critics will argue this is overreach — that banning nudifier systems won’t stop bad actors who can run open-source models locally. They’re not wrong. But the point of the law isn’t to eliminate every last instance of abuse; it’s to stop the commercial, scalable platforms from being the primary distribution channel for this filth. Grok, being a cloud-based service with millions of users, is exactly the kind of system that should be held accountable.

Musk’s habit of deflecting blame won’t fly here. The EU doesn’t care that users typed the prompts; it cares that the platform generated the output. And if the amendment passes the full Parliament — which seems likely given this vote — Grok and any similar service will have to either implement real content filters or face fines that make the GDPR penalties look like parking tickets.

This is higher than I expected, honestly. I figured the EU would take years to revisit the AI Act, not months. But when you have a high-profile case involving children and a billionaire CEO who’d rather blame his customers than fix his product, the political calculus shifts fast. Good.