Women are unclothed in photos by AI app due to the lack of legislation
By May Khong
On 2026 New Year’s Eve, Julie Yukari, a musician based in Brazil, posted a photo of herself on X.
At first, only her friends liked the post. The next morning, she found it had gone oddly viral, with over one thousand likes.
“Then I received a comment requesting Grok to put me in a microkini. By the time I realized, several users had requested nude photos of me from Grok, which it generated,” Yukari recalls.
“They looked disturbingly real with my face there. I felt ashamed, scared, and desperate,” she says.
Yukari’s manipulated images had been generated using Grok-2, X’s AI chatbot, after the platform introduced the new image-editing features in late 2025.
Elon Musk’s xAI launched Grok in November 2023, making the chatbot available directly on X.

Elon Musk’s generative AI chatbot Grok
By tagging “@grok” in public posts, users can prompt the tool to generate immediate responses, including “put this woman in see-through bikinis” or “remove her clothes”.
In December 2025, Elon Musk shared a photo generated by the chatbot of himself in a bikini on X.
This led to explosive interest in Grok’s image-editing tools, especially its ability to “remove” clothing from people.
Requests to sexualize images of real people soon flooded the platform.
According to The New York Times, Grok posted more than 4.4 million images in just nine days after Musk’s post, 41 per cent of which contained women nudity.
The images were sent to Yukari’s mother, a patient recovering from cancer surgery who was then distressed over her daughter’s reputation.
The musician even quit her band for fear that her videos would be used for pornographic content, and the toxic comments on X negatively affected her mental health.
“Many people said I was only complaining because I am fat. Because of this, my self-esteem has drastically changed. I started therapy to deal with these traumas,” she says.
Yukari had tried X’s reporting tools from the start, with inconsistent results. X had initially rejected her reports, stating that the users prompting Grok had not violated its rules.
“But days later, some images vanished without explanation. They pretended they never messed with my photos, some of which are still available online,” she says.
While many victims deleted their social media accounts to escape the harassment, Yukari decided to continue posting photos and discuss the incident.
“I felt like if I quit, I would be penalized for something others did to me. The goal of those sexualizing women and children online is to make us stop posting by threatening us with their perverted means. They think we should be submissive, but I am not giving them what they want,” she says.
Other victims have taken legal action.
As of October 2026, xAI faces multiple class-action lawsuits in the United States.
Similar measures were taken by two Dutch victim support groups in March, who successfully required the court to impose a daily penalty of €100,000 until X removes the nudity function.
The outcry prompted authorities worldwide to react.
Grok and X had been blocked altogether in Indonesia and Malaysia, although the ban was lifted afterwards.
The European Commission, Ireland’s Data Protection Commission, and Britain’s media regulator Ofcom all launched formal investigations into allegations of user-rights violations by X.
Under increasing pressure, X announced a series of measures on January 14, 2026. The platform stated it would geoblock all Grok users from creating intimate images in restricted regions.
In what is called an “extra layer of protection,” the photo editing tool would be put behind a paywall.
But for many users, the announcement offered little comfort.
User Mary* on X says her photos continued to be manipulated.
“Even with extra protection, people still found ways around it,” Mary says.
As an artist who actively promotes herself on X, Mary used to share her photos with a wide audience.
She first ignored the outcry against Grok, confident that the new measures would protect her privacy.
But after X’s January announcement, her photos were still edited and circulated on the platform.
“One follower, whose account is based in an unregulated area, requested Grok to put me in see-through clothes right under my post. I was horrified as the chatbot still spat out these images,” the artist says.
According to Reuters’ investigator team, the chatbot can still generate non-consensual nude images in certain circumstances. With the right prompts, Grok will sexualize photos of real people even without their consent, they reported.
To Mary, X should do more to protect users. “If you geoblock sensitive imagery in one country but allow it in another, what’s the meaning? It feels like X is just doing the bare minimum to cover up the scandal,” she says.
Unlike Yukari, Mary has lost faith in social media. She blocked all previous followers and locked her X account.
“Before, X used to be a comfort space, but now exposure makes me feel vulnerable. I admire people who continue speaking up online for the victims, but for those not brave enough like me, it’s best to retreat and stay away from posting photos as much as possible,” she says.
Assistant Professor Oliver Chan Ngai-keung of the School of Journalism and Communication at Tthe Chinese University of Hong Kong points out that Grok clearly violates users’ rights of dignity and privacy.
“Yet, many existing legal frameworks were not built for AI. Current laws, like Section 230 of the Communications Decency Act in the US, have shielded platforms from liability for user-generated content. Platforms can claim the right, rather than the responsibility, to moderate content and users’ interaction,s.” Chan says.
“User data and engagement are central to the business model of platforms and generative AI tools, so they don’t have a strong incentive to moderate content pro-actively. These firms strategically position themselves as ‘neutral’ intermediaries to avoid liability,” he adds.
The professor argues platform companies must conduct AI impact assessments before product launches and face punishments when prevention fails.
“Governments need to ensure that when harm occurs, legal enforcement carries real consequences on developers, rather than placing the burden on victims to navigate AI tools’ reporting processes,” he says.
In Hong Kong, the Personal Data Privacy Ordinance (Cap. 486) protects against the collection and misuse of personal data, including images. If a deepfake is created using someone’s photo without permission, the victim may file a complaint with evidence.
At present, Hong Kong has no law specifically targeting the creation of indecent images of another person, although existing laws may apply to related offences.
*Name changed at the request of interviewee
Edited by Suzana Li
Sub-edited by Justin Tai



































































