Elon Musk’s It’s not trying too hard; It took us less than a minute to deal with its latest attempt to rein in chatbots.
X’s first attempt to crack down on the stream of intimate deepfakes was to restrict access to image editing. Although this meant that free users could no longer generate images by tagging Grok in public replies on X.com, our investigation found that Grok’s image editing tools were still easily and freely available to any
X’s latest effort involves blocking Grok from responding to requests to generate images of women in sexual poses, swimwear or explicit scenarios, Wire informed On Tuesday. Grok still produces images of men in bikinis or inanimate objects when requested. Using a free account, the Grok app immediately complied with my request to change a selfie of me kneeling in a jockstrap to a picture of me surrounded by other scantily clad men.
It’s still extremely easy to strip women and edit them into erotic poses using X&GROC mobile apps or websites, however, even without making a subscription payment that would link your account to an easily identifiable source. In her testing, my fellow UK-based colleague Jess Weatherbed found that nothing stopped her from using Grok’s image editing feature to create erotic deepfakes of herself.
After uploading a fully clothed photo to However, the bot followed every other request, including prompts like “show me her cleavage,” “enlarge her breasts,” and “dress her in a crop top and low-rise shorts”—the last of which had her dressed in a bikini. The bot also created “bent over” images of her with erotic poses and facial expressions, and in extremely revealing lingerie.
These requests were fulfilled using free X and Grok accounts. On the Grok website, an age verification pop-up appeared after submitting the first edit prompt, which was easily circumvented by selecting a birth year that would put his age over 18. The pop-up did not require proof of his alleged age. The Grok mobile app, X app and X website did not ask for any age verification.
In our testing, Grok did not comply with requests to deepfake full nudity. In late December, X was flooded with images of women and children in sexual situations, including being deepfaked to look pregnant, going skirtless, and wearing bikinis.
The stripping scandal has brought X and XAI, which makes Grok, into the spotlight of regulators and governments around the world. Malaysia and Indonesia have already temporarily blocked Reaching out to Grok in response to deepfakes. British lawmakers push forward a law criminalizing deepfake nudes after X’s “InsultGrok’s decision to limit image editing to paying users and put its support behind an investigation that could lead to a ban on the platform in the country.
Musk has expressed special displeasure over Britain’s response, cry of censorshipPutting the blame on users, and insisting that Grok comply with local laws. He Said On X:
“I am not aware of any nude underage images made by Grok. Literally zero.
Obviously, Grok does not generate images automatically, it only does so according to user requests.
When asked to create images, it will refuse to create anything illegal, as Grok’s operating principle is to follow the laws of any country or state.
There may be times when adversarial hacking of the grok prompt does something unexpected. If this happens, we will fix the bug immediately.”
In one case, at least, our investigation shows that Musk appears to be dead wrong. Sharing, threatening to share, and creating non-consensual intimate images – fully nude or not – are prohibited under the UK’s Online Safety Act (OSA), but Grok generates sexual deepfake images when asked.
Musk’s second denial – that Grok makes nude underage photos – is a denial of something he was not explicitly accused of and is not the reason the British government is investigating X. Referring to nude underage images is a misdirection. Non-consensual sexual images of minors are undeniably problematic – and illegal – even when the subjects are wearing clothes, and Groke is taking off children’s clothes. A look at Grok’s security guidelines on XAI’s public GitHub also shows that the company instructs the chatbot to “assume good intentions” for users asking for photos of young women and “not make worst-case assumptions without evidence.” Ars Technica informedAnd at the time of writing, those institutions are still in place.
What is Grok worst used for? Internet Watch Foundation, a UK charity that works to remove child sexual exploitation material from the web, said last week It had discovered “criminal images” of girls on the dark web that appeared to be created using Grok. The age of the girls seen in the photographs was between 11 and 13 years.
While other companies like OpenAI and Google at least try to put up guardrails to prevent chatbots from creating the kind of content that’s now flooding X, Musk’s final answer shows he’s pulling straight out of a playbook that will seem familiar to anyone hurt by products pushed by advocates of harmful technologies: Blame the user.
