Grok image generation is now paywalled on X amid AI "undressing" controversy (update: not paywalled)

[midian182]

What just happened? With the pressure growing on X over Grok's sexualized deepfake images of women and children, the AI tool's image creation feature has now been restricted to paying X subscribers – but it can still be accessed by everyone through the standalone app. The move comes as British Prime Minister Keir Starmer suggested that X be blocked in the UK.

There has been growing outcry over Grok after it was revealed that the chatbot is being used to generated nude and other sexualized deepfakes, sometimes involving minors. A study found that it created about 6,700 images every hour that were identified as sexually suggestive or nudifying. For comparison, the other top websites for such content average 79 similar images per hour combined.

Update (12pm): As verified by multiple users, unsubscribed X users can still edit images, with the platform only limiting requests when publicly replying to the Grok chatbot. Image editing remains available on the desktop site, while app users can access the same feature by long-pressing on an image. By avoiding public prompts, X's intent seems to leave controversial image edits out of the public feed as a rapid fix.

With numerous countries investigating the matter, the UK considering a ban on X, and the company continuing to blame users rather than Grok, restrictions have now been introduced.

"Image generation and editing are currently limited to paying subscribers," Grok wrote in an X post. The change means that only people who have their full details and credit card information stored on X's systems can use Grok to create images directly on X. It's presumed that as they can be identified, subscribers won't create anything they shouldn't.

The caveat is that anyone – non-paying users included – can still generate images on the separate standalone Grok app, which does not share the images it creates publicly.

1960: in 2026 we'll have flying cars

2026: "hey grok put this guy in a bikini" pic.twitter.com/FEccHnIsWX

– Tiffany Fong (@TiffanyFong) January 4, 2026

The Internet Watch Foundation (IWF) has confirmed that Grok had been used to create "criminal imagery of children aged between 11 and 13."

Elon Musk previously said that anyone using Grok to create illegal content would face the same consequences as having uploaded such material directly.

The UK government has reacted to the new restriction. It condemned the move as simply making the ability to generate explicit and unlawful images a premium service.

"It's not a solution. In fact, it's insulting to victims of misogyny and sexual violence. What it does prove is that X can move swiftly when it wants to do so. You heard the prime minister yesterday. He was abundantly clear that X needs to act, and needs to act now. It is time for X to grip this issue," a Downing Street spokesperson said.

The spokesperson added that it would support any action taken by Ofcom, the UK's media regulator.

In a reply to users last week, Grok said that most cases of minors appearing in its generated sexualized images could be prevented through advanced filters and monitoring, but it admitted that "no system is 100% foolproof." It added that xAI was prioritizing improvements and reviewing details shared by users.

Permalink to story:

Grok image generation is now paywalled on X amid AI "undressing" controversy (update: not paywalled)

 

[Kirby1]

So now people have to pay a few buck to generate child porn. Problem solved.

Any Musk fans starting to feel silly yet?

 

[PXAbstraction]

So now people have to pay a few buck to generate child porn. Problem solved.

Any Musk fans starting to feel silly yet?

If anything, that makes X AI look even more liable because now they're explicitly generating income from CSAM content.

 

[wiyosaya]

If anything, that makes X AI look even more liable because now they're explicitly generating income from CSAM content.

Exactly! Much like SCOTUS has determined that "Free Speech rights" do not include inciting violence, making people pay to create porn will not eschew GROK culpability.

fElon is a dolt who is beholden to his fake image as the "worlds smartest man" even though wealth /= brains.

 

[ScottSoapbox]

The fact that there is a tracked AI comparison metric of sexualized images generated per hour shows just quickly we’re going down a dark rabbit hole.

 

[unoficialoficial]

Generative(no real human authorship) pixels do not come directly from real events with real participants, so they cannot produce a violation of rights.

 

[PXAbstraction]

Generative(no real human authorship) pixels do not come directly from real events with real participants, so they cannot produce a violation of rights.

That's an interesting take. I'm really curious how that would hold up in court, genuinely.

 

[sreams]

I don't know... banning AI tools over this seems the same as banning pencils and pens. Anybody can draw something offensive using a tool and then post it for all to see. If that posting violates a law, the person posting/drawing it should be the target for legal action.

 

[PXAbstraction]

I don't know... banning AI tools over this seems the same as banning pencils and pens. Anybody can draw something offensive using a tool and then post it for all to see. If that posting violates a law, the person posting/drawing it should be the target for legal action.

Due respect, I think that's a false equivalence, similar to when people say if you ban guns, you should also ban knives because they can both technically kill people. Pencils can't generate 10s of these types of images per minute, in exquisite detail. Pencils require someone to have actual artistic skill to make things with them, especially if they're going to make something that resembles someone else. Taking that drawing you made with the pencil and posting it to places with significant reach isn't easy. You think if I hand drew CSAM and posted it to Twitter, that I wouldn't get banned in minutes?

There's a reason you can't easily do this with other AI models and it's because they designed them to not be able to. That Grok is able to, that they haven't done anything to stop it and now are in fact paywalling it is very telling about their ethics and motivations.

 

[sreams]

Due respect, I think that's a false equivalence, similar to when people say if you ban guns, you should also ban knives because they can both technically kill people. Pencils can't generate 10s of these types of images per minute, in exquisite detail. Pencils require someone to have actual artistic skill to make things with them, especially if they're going to make something that resembles someone else. Taking that drawing you made with the pencil and posting it to places with significant reach isn't easy. You think if I hand drew CSAM and posted it to Twitter, that I wouldn't get banned in minutes?

There's a reason you can't easily do this with other AI models and it's because they designed them to not be able to. That Grok is able to, that they haven't done anything to stop it and now are in fact paywalling it is very telling about their ethics and motivations.

AI doesn't have a will. These images come into existence only when a human wants them to and specifically requests/describes them. They aren't responsible for their actions? Have any of the people making these images been arrested?

 

[PXAbstraction]

AI doesn't have a will. These images come into existence only when a human wants them to and specifically requests/describes them. They aren't responsible for their actions? Have any of the people making these images been arrested?

You're shifting the goalposts. Yes, the people made the images, using Grok. They didn't make them using ChatGPT or Copilot or Gemini because those tools have safeguards that prevent using them for CSAM. Most of the people making these images haven't been arrested because Grok also have no safeguards to make them easily identifiable. Sure, you have to have a paid account now, but it's not like you have to do identity verification before you sign up.

Why you're continuing to defend Musk's ability to profit off of CSAM that even the other sociopathic AI CEOs won't allow is confusing to me.

 

[sreams]

You're shifting the goalposts. Yes, the people made the images, using Grok. They didn't make them using ChatGPT or Copilot or Gemini because those tools have safeguards that prevent using them for CSAM. Most of the people making these images haven't been arrested because Grok also have no safeguards to make them easily identifiable. Sure, you have to have a paid account now, but it's not like you have to do identity verification before you sign up.

Why you're continuing to defend Musk's ability to profit off of CSAM that even the other sociopathic AI CEOs won't allow is confusing to me.

If we removed tools based on the concern that someone might profit from them, we wouldn't have any tools. If authorities have difficulty identifying someone who commits a crime, they should get a subpeona and get whatever information they need from X or any of the myriad other companies that offer tools.

 

[PXAbstraction]

If we removed tools based on the concern that someone might profit from them, we wouldn't have any tools. If authorities have difficulty identifying someone who commits a crime, they should get a subpeona and get whatever information they need from X or any of the myriad other companies that offer tools.

You do realize you're trying to defend a tool lack of restrictions causing it to be used to generate massive amounts of CSAM and non-consentual pornographic materials...right? No one's saying Grok should remove image generation tools (at least I'm not), I'm saying they should put the same restrictions on them that all their competitors have that prevent them for being used for those things. This is already a solved problem, they just aren't also solving it. Please tell me you see a need for them to be accountable for that.