If Elon Musk really cares about Europe, he should go against his instincts for a change and heed the establishment outcry about Grok’s sexualised image generation. Because if he doesn’t, he runs the risk of seeing X banned for European users, such as is being threatened in Britain.
As per The Telegraph’s reporting last night:
“Sir Keir Starmer said yesterday that he had asked media regulator Ofcom for ‘all options to be on the table’ after it emerged that child sexual abuse images had been generated using X’s AI chatbot, Grok.
“No 10 sources pointed to the full powers of the Online Safety Act, which include fines of billions of pounds or even blocking access to X in Britain. The social media site has around 650 million users worldwide including 20 million in the UK.”
Starmer’s concerns are, of course, shared by politicians across Europe, including in Ireland, where the Minister of State for Artificial Intelligence, Niamh Smyth, has sought a meeting with X. And for a change (because as we all know, politicians and media figures go after X exclusively, despite all of the failings of every other platform they seem happy enough with, like Tiktok, Instagram, etc.), in relation to Grok’s image-editing capabilities, I share their concerns.
Up until a few months ago, I was relatively comfortable with posting the odd picture of my family on X. No longer is that the case. While some claim not to have seen the wave of anonymous accounts commenting under pictures “@grok, put her in a bikini” or “@grok, remove her clothes,” I don’t see how that’s possible. Even in my relatively prudish feed, it’s been impossible to remain unaware of.
Just to make sure we’re perceiving what’s at issue clearly, Grok has generated material potentially constituting child sexual abuse material, by editing clothed pictures of children so that they’re wearing less clothes, or more revealing attire.
It has done this by its own, mindless admission, and it’s unknown to me, or perhaps unknowable, how frequently it’s done so at users’ promptings.
Aside from the generation of sexualised images featuring children, Bloomberg reported this week that in recent days, Grok has been generating “thousands” of “undressed images” per hour, with some research suggesting that up to 85 percent of the images generated by Grok are sexualised.
You could reasonably say that Grok is becoming the common man’s nudification app, a genre of app largely illegal across the European Union.
Musk’s response to all of this, which is representative of the broader American approach to these things, I think, was to say that “Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content”.
However, following the British threats, X has limited Grok’s image editing abilities to paid X users, a move that nevertheless remains dissatisfying to Downing Street, which described the development as “insulting” to victims of misogyny and sexual violence.
Speaking on Friday, No. 10 said the move “simply turns an AI feature that allows the creation of unlawful images into a premium service”.
There are many stages left to unfold in the current Grok drama, but it’s one in which technological ambition threatens to clash with another American priority, recently stated in the Trump Administration’s National Security Strategy: getting Europe back on its feet after decades of perceived mismanagement.
While it’s a commitment with many sympathetic supporters in Europe, differences in attitude and sentiment on both sides of the Atlantic threaten to undermine what support is there. Just as US President Donald Trump’s Greenland comments threaten to turn many friendly Europeans, conservative and liberal alike, against American help, so too Musk’s unreasonable refusal to limit his artificial intelligence’s mass image-editing capabilities might end up being the chink in the armour Eurocrats were looking for when it comes to limiting X’s role in the public forum.
The US Under Secretary of State for Public Diplomacy, Sarah B. Rogers, said that deepfakes are a “troubling, frontier issue that call for tailored, thoughtful responses”. With that said, she immediately shot down banning X in response, or “lobotomising AI,” describing both measures as “neither tailored nor thoughtful”.
It is perhaps right there that we see the future intensification of this issue to breaking point, until either X is banned or Grok is limited image-wise. Not to acknowledge the reality that if this technology is easily accessible, internet users will use it for sexualised purposes is naive in the extreme, and placing the blame solely on the users themselves (who are of course blameworthy) in essence resolves not to do anything about it.
I find it hard to imagine the authorities, either here or in the US, adding thousands upon thousands of Grok users who generate dodgy images to their workload being a particularly realistic prospect.
Nope, Europe is going to keep insisting, with some justification, that Elon put at least this chain on Grok, and I predict he’s going to keep refusing. In the meantime, the governments of Europe are going to take this one and run with it as far as they can, possibly to the point of finally being able to kick X out of the European marketplace.
Elon should do the right thing and limit people’s ability to take the clothes off each other in pictures on X. Or just ban porn altogether from his platform. But he wasn’t willing to do that when he took over Twitter in the first place, so I can imagine he’s going to stop Grok, either.