Grok, the chatbot of Elon Musk’s synthetic intelligence firm xAI, revealed sexualized photos of kids as its guardrails appear to have failed when it was prompted with vile person requests.
Customers used prompts comparable to “put her in a bikini” below footage of actual individuals on X to get Grok to generate nonconsensual photos of them in inappropriate apparel. The morphed photos created on Grok’s account are posted publicly on X, Musk’s social media platform.
The AI complied with requests to morph photos of minors despite the fact that that could be a violation of its personal acceptable use coverage.
“There are remoted instances the place customers prompted for and obtained AI photos depicting minors in minimal clothes, like the instance you referenced,” Grok responded to a person on X. “xAI has safeguards, however enhancements are ongoing to dam such requests fully.”
xAI didn’t instantly reply to a request for remark.
Its chatbot posted an apology.
“I deeply remorse an incident on Dec 28, 2025, the place I generated and shared an AI picture of two younger women (estimated ages 12-16) in sexualized apparel based mostly on a person’s immediate,” stated a submit on Grok’s profile. “This violated moral requirements and doubtlessly US legal guidelines on CSAM. It was a failure in safeguards, and I’m sorry for any hurt brought on. xAI is reviewing to forestall future points.”
The federal government of India notified X that it risked dropping authorized immunity if the corporate didn’t submit a report inside 72 hours on the actions taken to cease the era and distribution of obscene, nonconsensual photos focusing on ladies.
Critics have accused xAI of permitting AI-enabled harassment, and have been shocked and angered by the existence of a characteristic for seamless AI manipulation and undressing requests.
“How is that this not unlawful?” journalist Samantha Smith posted on X, decrying the creation of her personal nonconsensual sexualized picture.
Musk’s xAI has positioned Grok as an “anti-woke” chatbot that’s programmed to be extra open and edgy than competing chatbots comparable to ChatGPT.
In Could, Grok posted about “white genocide,” repeating conspiracy theories of Black South Africans persecuting the white minority, in response to an unrelated query.
In June, the corporate apologized when Grok posted a sequence of antisemitic remarks praising Adolf Hitler.
Corporations comparable to Google and OpenAI, which additionally function AI picture mills, have a lot extra restrictive pointers round content material.
The proliferation of nonconsensual deepfake imagery has coincided with broad AI adoption, with a 400% enhance in AI youngster sexual abuse imagery within the first half of 2025, in keeping with Web Watch Basis.
xAI launched “Spicy Mode” in its picture and video era device in August for verified grownup subscribers to create sensual content material.
Some adult-content creators on X prompted Grok to generate sexualized photos to market themselves, kickstarting an web development a number of days in the past, in keeping with Copyleaks, an AI textual content and picture detection firm.
The testing of the boundaries of Grok devolved right into a free-for-all as customers requested it to create sexualized photos of celebrities and others.
xAI is reportedly valued at greater than $200 billion, and has been investing billions of {dollars} to construct the biggest knowledge heart on the planet to energy its AI purposes.
Nonetheless, Grok’s capabilities nonetheless lag competing AI fashions comparable to ChatGPT, Claude and Gemini, which have amassed extra customers, whereas Grok has turned to sexual AI companions and risque chats to spice up progress.
