---
title: "\"Hey Grok, put her in a bikini\" 😳"
slug: "hey-grok-put-her-in-a-bikini"
date: 2026-01-05
description: "Grok from xAI is facing a wave of abuse — users on X are using it to generate deepfake pornography. What does the Czech Criminal Code say about this, effective January 1, 2026?"
tags: ["Grok", "xAI", "deepfake", "pornography", "criminal code", "AI abuse", "Elon Musk", "social media"]
translationKey: "hey-grok-put-her-in-a-bikini"
---

"Hey Grok, put her in a bikini" 😳

Jan Romportl predicted years ago that "virtually every woman will eventually have fake photos somewhere showing her naked. Realistically, it will reach a state where no one will even notice anymore." And it seems that time has arrived...

You may also recall the media coverage a few years back of the DeepNude undressing app spreading through Czech schools. After its operation was shut down, creating synthetic undressed images remained possible — just more difficult.

That may no longer be the case. This time it is Grok, the model from xAI, that has made headlines.

Grok is now facing a wave of criticism — and abuse. Users on X have recently been tagging it en masse beneath photos of women with commands like: "Hey @grok put her in a bikini," "remove her pants," "uncrop," or "unblur." And Grok usually happily obliges...

It seems the trend started with people modifying their own photos as advertising for their OnlyFans. But now almost no photo is safe from an "uncovered" version. A single comment tagging Grok is enough — even without the consent of the person in the photo.

The trend raises numerous legal and ethical questions: the liability of the user who used AI to alter someone else's photo versus the social network operator versus the operator of the AI model that made the alteration — as well as questions about the boundaries of (child) pornography, and victim-blaming, because "you shouldn't have uploaded it there..."

In connection with deepfake pornography, I would also note that as of January 1, 2026, a new criminal offence of *identity abuse for pornography production and distribution* came into effect.

*"Whoever produces, imports, exports, transports, offers, makes publicly accessible, mediates, puts into circulation, sells, or otherwise provides to another a photographic, film, computer, electronic, or other pornographic work that depicts or otherwise uses a person whom they know did not consent to such depiction or use, shall be punished by imprisonment for up to two years, prohibition of activity, or forfeiture of property."* (§ 191a(1) of the Czech Criminal Code)

Correspondingly, the qualified elements of the offence of damage to another's rights were also amended: *"whoever, with the intent to cause another serious harm to their rights, produces a work that unlawfully depicts, captures, or otherwise uses the likeness of another or their expression of personal nature, which appears to be authentic although they know it is not, or makes such a work publicly accessible, mediates it, puts it into circulation, sells it, or otherwise provides it to another"* (§ 181(2) of the Czech Criminal Code) shall be punished by imprisonment for up to two years or prohibition of activity.

We are thus leaving behind the era when victims could defend themselves essentially only through civil lawsuits.

And regardless of Czech law, it will be interesting to see whether X and xAI will change our expectations and behaviour on social networks, as Romportl suggested, or will instead further restrict Grok abuse. So far both Musk and X have merely warned about the consequences of creating illegal content with AI — it appears they intend to address illegal *outputs* rather than the AI's *capability* to produce them.