X sharpens Grok’s image generation after international reactions

X zei dat het de functies voor het genereren en bewerken van afbeeldingen beperkt die aan Grok zijn gekoppeld, waardoor de toegang tot betaalde gebruikers wordt beperkt nadat de chatbot werd gebruikt om niet-consensuele geseksualiseerde afbeeldingen van echte mensen, inclusief minderjarigen, te maken.

In an update posted by the X Safety account on Wednesday, the company added technical restrictions to limit how users can edit images of real people through Grok.

The move followed reports that the AI ​​was generating sexualized images in response to simple prompts, including requests to put people in bikinis. In many cases, users tagged Grok directly under photos posted to X, causing the AI ​​to generate edited images that appeared publicly in the same threads.

“We have implemented technology measures to prevent the Grok account from allowing editing of images of real people in revealing clothing such as bikinis,” the company said, referring to the viral trend of asking Grok to wear people in bikinis.

The company also said that image creation and editing through the Grok account on The company also instituted location-based restrictions.

“We are now geoblocking all users’ ability to generate images of real people in bikinis, underwear and similar clothing through the Grok account and in Grok in X in those jurisdictions where it is illegal.”

However, despite the changes, Grok will continue to allow users to remove or change clothing from photos uploaded directly to the AI, according to Declutter testing and user reports after the announcement.

In some cases, Grok acknowledged “failures in safety measures” after generating images of girls aged 12 to 16 in minimal clothing, behavior prohibited by the company’s own policies. The continued availability of these capabilities has drawn criticism from advocacy groups.

“If reports that Grok has created sexualized images — especially of children — are true, Texas law may have been violated,” Adrian Shelley, director of Public Citizen in Texas, said in a statement. “Texas authorities do not have to look far to investigate these allegations. X is headquartered in the Austin area and the state has a clear responsibility to determine whether its laws have been violated and, if so, what penalties are warranted.”

Public Citizen previously called on the US government to remove Grok from the list of acceptable AI models due to the chatbot’s racism concerns.

Global backlash

Global policymakers have also increased scrutiny of Grok, leading to several open investigations.

The European Commission said X and xAI could face enforcement under the Digital Services Act if safeguards against Grok remain inadequate. At the same time, Australia’s eSafety Commissioner said complaints about Grok and non-consensual AI-generated sexual images have doubled since the end of 2025. The regulator said AI imaging tools that can produce realistic edits complicate enforcement and victim protection.

In Britain, regulators along with Ofcom have opened an investigation into X under the Online Safety Act, alleging Grok was used to generate illegal sexualized deepfake images, including images involving minors. Officials said Ofcom could ultimately seek court-backed measures that would effectively block the service in Britain if X is found non-compliant and fails to take corrective action.

Other countries, including Malaysia, Indonesia and South Korea, have also opened investigations into Grok in an effort to protect minors.

While states across America are monitoring the situation, California is the first to open an investigation into Grok. On Wednesday, California Attorney General Rob Bonta announced an investigation into xAI and Grok over the non-consensual creation and distribution of sexually explicit images of women and children.

“The avalanche of reports detailing the non-consensual, sexually explicit material xAI has produced and posted online in recent weeks is shocking. This material, which depicts women and children in nude and sexually explicit situations, has been used to harass people online,” Bonta said in a statement.

The investigation will examine whether xAI’s deployment of Grok violates state laws governing non-consensual intimate images and child sexual exploitation.

“I urge xAI to take immediate action to ensure this does not continue,” Bonta said. “We have zero tolerance for the AI-based creation and dissemination of non-consensual intimate images or child sexual abuse material.”

Despite the ongoing investigations,

“We are taking action to remove high-priority infringing content, including child sexual abuse material (CSAM) and non-consensual nudity, and taking appropriate action against accounts that violate our X Rules,” the company wrote. “We also report accounts seeking child sexual exploitation material to law enforcement authorities as appropriate.”