The ability to generate AI images using Elon Musk's chatbot, Grok, has now been restricted to paying users, following a worldwide regulatory backlash and public outrage at the bot producing thousands of non-consensual, sexualised AI deepfakes.
Grok is operated by xAI, the Elon Musk-led parent company of X (formerly Twitter).
Since introducing the ability to edit images in late December, people have been using Grok, accessible via the social media platform, to remove people's clothes through the chatbot's "spicy" mode.
That feature has been flooding the platform with sexualised deepfakes, mainly of women, and potentially producing nudified images of minors.
Now, on Friday (9 January), when a free-tier X user asks for an edited image, Grok responds: “Image generation and editing are currently limited to paying subscribers. You can subscribe to unlock these features.” And with a link to subscribe.
To access the image editor via the social media platform, a one-month subscription costs €9.39.
Reviewing the chatbots' replies on the X on Friday, sexualised deepfakes by paying users are still being generated and publicly visible. Though there has been a seeming decline in sexualised Grok replies from previous days.
The change happened as multiple world governments, plus the European Commission, denounced such images and began looking into the chatbot's behaviour.
Responding to the update, a commission spokesperson said to the press on Friday: “We, of course, have taken note of the recent changes announced by X for Grok. Indeed, limiting the image generation and editing to paying subscribers".
"This doesn't change our fundamental issue. Paid subscription or non-paid subscription, we don't want to see such images. It's as simple as that.”
This incident has raised legal questions about who is liable for these images, with Irish media minister Patrick O’Donovan telling the media that the individual prompters are ultimately responsible. At the same time, experts have argued that Musk and xAI are also culpable.
It also raises questions about what EU regulation can do about the potentially illegal content.
Grok has produced images that could seemingly be considered child sexual abuse material.
Researchers at an EU-based non-profit AI Forensics investigated Grok, and found, over the Christmas and New Year's period, thousands of sexualised deepfakes, and 30 images depicting young women and girls in minimal clothing.
Commenting to EUobserver on the more recent developments, AI Forensics said it was pleased the commission is taking the matter seriously.
Researcher Paul Bouchaud said: "We note that the platform has taken some minor action, restricting the generation to paid X users, while leaving previously generated images available on the platform.”
“Furthermore, we continue to observe the generation of sexualised images on X and full pornographic videos via the Grok Imagine feature," Bouchaud added.
According to an analysis by technology news outlet TechCrunch, they estimate X had roughly 1.4 million paying users in 2024.
Become a subscriber and support EUobserver's journalism in 2026.
Owen Carpenter-Zehe is a junior reporter from the US at EUobserver, covering European politics.
Owen Carpenter-Zehe is a junior reporter from the US at EUobserver, covering European politics.