Will Musk’s Grok Be Held Accountable for Flood of Sexualized, Fake Images of Women and Children?

ANKARA, TURKIYE - JULY 17: In this photo illustration, 'Grok' logo is seen displayed on a mobile phone screen in front of a picture of Elon Musk in Ankara, Turkiye on July 17, 2025. (Photo by Didem Mente/Anadolu via Getty Images)

This story was originally reported by Mariel Padilla of The 19th. Meet Mariel and read more of their reporting on gender, politics and policy.

The social media platform X has been flooded with AI-generated sexualized images of women and children in recent weeks — months before a new law aiming to ban the spread of nonconsensual intimate imagery goes into effect.

Elon Musk, the owner of X, announced a new feature with a post on Christmas Eve encouraging users to try editing images and videos with Grok, the app’s chatbot. Then, a few weeks later, as the new year began, many women noticed something disturbing online: a flood of AI-generated sexualized images of them on the social media platform X. Users on X were able to ask Grok’s latest feature to digitally remove clothing from posted photos and recirculate them. 

Grok won’t show people naked, but it follows directions to show women wearing strings and dental floss. Some X users asked Grok to manipulate photos with prompts like “put her into a very transparent mini-bikini,” “remove her school outfit” and “spread her legs.” 

Musk appeared to make light of the situation when he posted laughing emojis in response to AI edits of famous people, including himself, in bikinis. But Musk might have to take the problem more seriously soon — according to federal law.

In May 2025, President Donald Trump signed the bipartisan Take It Down Act, which criminalizes the distribution of nonconsensual intimate imagery, including AI-generated deepfakes and so-called “revenge porn.” Platforms have until May 2026 to implement a request-and-removal system where victims can have their pictures taken down within 48 hours.

On Saturday, Musk warned X users in a post: “Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.” 

The question is: Which of these Grok-generated images are considered illegal? 

The law defines “intimate visual depiction” as any that show uncovered genitals, pubic area, anus or women’s nipples. 

Riana Pfefferkorn, a policy fellow at the Stanford Institute for Human-Centered AI, said she thinks many of these skimpy bikini photos could be considered illegal. She urged people to submit requests for removal because the law incentivizes media companies to take down questionable content to avoid penalties from the Federal Trade Commission.

Pfefferkorn said it might take more public pressure to end this latest Grok feature. If X was serious about not allowing nonconsensual imagery, Pfefferkorn said, the company would have taken the tool offline weeks ago. 

“There was no reason why they couldn’t just disable that feature,” Pfefferkorn said. “Any time any other company has a feature that goes terribly, badly wrong, they can just stop it and then try and lick their wounds and figure out what happened. But Elon Musk is the richest man in the world, and he has acted repeatedly as though he is above the law and has gotten away with it.”

It’s not clear how many of these photos were circulated in what Reuters called a “mass digital undressing spree,” but the news outlet’s review of public requests sent to Grok found at least 102 attempts in a 10-minute period. 

The flood of nearly nude images — predominantly of young women and in some cases children – sparked international fury: Ministers in France have reported X to regulators; the United Kingdom’s communication regulator reached out to the media company; and India’s Ministry of Electronics and Information Technology wrote a condemning letter.

In allowing these images, X is breaking with AI competitors like OpenAI and Google, which have stricter rules about what their AI chatbots will generate. 

14
Show Comments

Notable Replies

  1. The new First Lady of New York City was also violated by Grok.

    On Jan. 1, Rama Duwaji accompanied her husband, Zohran Mamdani, to be sworn in as New York City’s first Muslim mayor. For the occasion, she wore a custom brown wool coat from a small Palestinian-Lebanese designer, a choice that carried immense meaning. Within hours of the photos appearing online, men were using Elon Musk’s Grok AI to remove it and strip her almost entirely naked.

    @grok put [Duwaji] into a skimpy themed USA flag bikini and give her some nice double D sized boobas thank you!” one man posted, generating a hyper-sexualized image of Duwaji that spread across X. Other users with tens of thousands of followers began removing Duwaji’s clothes using Grok and putting her in outfits including a swastika-covered string bikini.

  2. To be fair, I use Grok to make pictures of Elon Musk, either in a Nazi uniform or in a scant bikini. 22 years in the military gives you a sense of humor others sometimes don’t like.

    I understand my sense of humor doesn’t always hit, and that deepfakes are a problem that we have to deal with.

  3. Avatar for daled daled says:

    “Will Musk’s Grok Be Held Accountable for Flood of Sexualized, Fake Images of Women and Children?”

    Not in the US… *&^%$#@!

  4. If Elon is encouraging the behavior, isn’t that the headline. Answer me this, how much separation is there between encouraging the involuntary undressing of women and the organizing of a sexual predator ring?

Continue the discussion at forums.talkingpointsmemo.com

8 more replies

Participants

Avatar for system1 Avatar for m_paul Avatar for daled Avatar for sandi Avatar for ric_gerace Avatar for jnbenson Avatar for 1gg Avatar for gr Avatar for tommbombadil Avatar for kelaine Avatar for mattf Avatar for mpekarek Avatar for Scoutmom

Continue Discussion