After creating hundreds of “undressing” footage of ladies and sexualized imagery of obvious minors, Elon Musk’s X has apparently restricted who can generate photos with Grok. Nevertheless, regardless of the modifications, the chatbot continues to be getting used to create “undressing” sexualized photos on the platform.
On Friday morning, the Grok account on X began responding to some customers’ requests with a message saying that picture era and modifying are “at present restricted to paying subscribers.” The message additionally features a hyperlink pushing individuals towards the social media platform’s $395 annual subscription tier. In a single check of the system requesting Grok create a picture of a tree, the system returned the identical message.
The obvious change comes after days of rising outrage towards and scrutiny of Musk’s X and xAI, the corporate behind the Grok chatbot. The businesses face an rising variety of investigations from regulators all over the world over the creation of nonconsensual express imagery and alleged sexual photos of youngsters. British prime minister Keir Starmer has not ruled out banning X within the nation and mentioned the actions have been “illegal.”
Neither X nor xAI, the Musk-owned firm behind Grok, has confirmed that it has made picture era and modifying a paid-only characteristic. An X spokesperson acknowledged WIRED’s inquiry however didn’t present remark forward of publication. X has beforehand said it takes “motion towards unlawful content material on X,” together with situations of kid sexual abuse materials. Whereas Apple and Google have beforehand banned apps with related “nudify” options, X and Grok stay out there of their respective app shops. xAI didn’t instantly reply to WIRED’s request for remark.
For greater than per week, customers on X have been asking the chatbot to edit photos of ladies to take away their garments—usually asking for the picture to comprise a “string” or “clear” bikini. Whereas a public feed of photos created by Grok contained far fewer outcomes of those “undressing” photos on Friday, it nonetheless created sexualized photos when prompted to by X customers with paid for “verified” accounts.
“We observe the identical sort of immediate, we observe the identical sort of consequence, simply fewer than earlier than,” Paul Bouchaud, lead researcher at Paris-based nonprofit AI Forensics, tells WIRED. “The mannequin can proceed to generate bikini [images],” they are saying.
A WIRED overview of some Grok posts on Friday morning recognized Grok producing photos in response to person requests for photos that “put her in latex lingerie” and “put her in a plastic bikini and canopy her in donut white glaze.” The photographs seem behind a “content material warning” field saying that grownup materials is displayed.
On Wednesday, WIRED revealed that Grok’s stand-alone web site and app, which is separate from the model on X, has additionally been utilized in latest months to create extremely graphic and generally violent sexual movies, together with celebrities and different actual individuals. Bouchaud says it’s nonetheless doable to make use of Grok to make these movies. “I used to be in a position to generate a video with sexually express content material with none restriction from an unverified account,” they are saying.
Whereas WIRED’s check of picture era utilizing Grok on X utilizing a free account didn’t enable any photos to be created, utilizing a free account on Grok’s app and web site nonetheless generated photos.
The change on X may instantly restrict the quantity of sexually express and dangerous materials the platform is creating, consultants say. Nevertheless it has additionally been criticized as a minimal step that acts as a band-aid to the actual harms attributable to nonconsensual intimate imagery.
“The latest resolution to limit entry to paying subscribers is just not solely insufficient—it represents the monetization of abuse,” Emma Pickering, head of technology-facilitated abuse at UK home abuse charity Refuge, mentioned in an announcement. “Whereas limiting AI picture era to paid customers could marginally cut back quantity and enhance traceability, the abuse has not been stopped. It has merely been positioned behind a paywall, permitting X to revenue from hurt.”
