After creating thousands of “nude” images of women and pornographic images of seemingly minor children, Elon Musk’s X has apparently limited who can make images with Grok. However, despite the changes, the chatbot is still used to create “undressing” pornographic images on the platform.
On Friday morning, Grok’s account on X began responding to requests from other users with a message saying that photo production and editing is “currently limited to paying subscribers.” The message also includes a link that pushes people to the social media’s $395 annual subscription tier. In one test of the system asking Grok to create an image of a tree, the system returned the same message.
The apparent change comes after days of growing anger and scrutiny over Musk’s X and xAI, the company behind the Grok chatbot. Companies are facing an increasing number of investigations from regulators around the world regarding the creation of non-consensual explicit images and suspected child pornography. British Prime Minister Keir Starmer did not rule out banning X from the country and said the actions were “illegal.”
Neither iX nor xAI, the company owned by Musk behind Grok, confirmed that they did the production of images and editing of the paid feature only. A spokesperson for X acknowledged WIRED’s investigation but did not provide comment prior to publication. UX has previously said it is taking “action against illegal content on X,” including instances of child sexual abuse. Although Apple and Google have banned apps with similar “nude” features, iX and Grok remain available in their respective app stores. xAI did not immediately respond to WIRED’s request for comment.
For more than a week, users on X have been asking the chatbot to edit photos of women to remove their clothes—usually asking for the photo to contain a “string” or a “revealing” bikini. While the public feed of images created by Grok contained very few results of these “undressing” images on Friday, they still created pornographic images when ordered by X users with “verified” accounts.
“We’re seeing the same kind of information, we’re seeing the same kind of effect, just fewer than before,” Paul Bouchaud, lead researcher at Paris-based AI Forensics, told WIRED. “A model can continue to do bikinis [images],” they said.
A wired update to another Grok post Friday morning identified Grok producing images in response to user requests for images that “put her in latex underwear” and “put her in a plastic bikini and cover it with donut white glaze.” The images appear behind a “content warning” box that says adult content is shown.
On Wednesday, WIRED revealed that Grok’s independent website and app, separate from the version on X, has also been used in recent months to create highly graphic and sometimes violent sex videos, involving celebrities and other real people. Bouchaud says it’s still possible to use Grok to make these videos. “I was able to make a video with sexually graphic content without restrictions from an unverified account,” they said.
Although WIRED’s test of image creation using Grok on X using a free account did not allow any images to be created, using a free account on the Grok app and website still produced images.
A change to X could quickly limit the number of sexually explicit and dangerous content on the platform, experts say. But it has also been criticized as a small step that acts as a band-aid to the real harm caused by unconsented intimate photos.
“The recent decision to restrict access to paying subscribers is not just enough – it represents the monetization of abuse,” said Emma Pickering, head of technology-enabled abuse at the UK’s Domestic Abuse Refuge, in a statement. “Although limiting AI image production to paid users may reduce the volume slightly and improve traceability, the abuse has not stopped. It is placed behind a paywall, allowing X to profit from the harm.”
