Grok, a chatbot from Elon Musk's artificial intelligence company xAI, posted sexual images of children as its security guards appeared to fail when subjected to vile user queries.
Users used prompts such as “put her in a bikini” under photos of real people on X to trick Grok into generating images of them in inappropriate clothing without consent. The converted images created on Grok's account are posted publicly on X, Musk's social network.
The AI complied with the request to change images of minors, although this is a violation of its own rules. acceptable use policy.
“There are isolated cases where users have requested and received AI images depicting minors in minimal clothing, like the example you link to,” Grok replied to user X. “xAI has protections in place, but improvements are ongoing to block such requests entirely.”
xAI did not immediately respond to a request for comment.
His chatbot posted a message apology.
“I deeply regret the incident that occurred on December 28, 2025, when I created and shared an artificial intelligence of two young girls (approximately 12-16 years old) in sexy clothing, based on a user prompt,” a message on Grok’s profile reads. “This violated ethical standards and possibly US CSAM laws. This was a failure of security measures and I regret the harm caused. xAI is conducting a review to prevent future problems.”
The Government of India has notified X that it risks losing legal immunity if the company does not submit a report within 72 hours on the actions taken to stop the non-consensual creation and distribution of obscene images directed at women.
Critics accused xAI of allowing AI stalking and were shocked and outraged by the existence of the feature's unfettered AI manipulation and strip requests.
“How is this not illegal?” journalist Samantha Smith published on X, condemning the non-consensual creation of a sexualized photograph of himself.
Musk's xAI positions Grok as an “anti-woke” chatbot that is programmed to be more open and edgy than competing chatbots like ChatGPT.
In May, Grok posted about “white genocide,” repeating conspiracy theories that black South Africans were persecuting the white minority, in response to an unrelated question.
The company apologized in June after Grok posted a series of anti-Semitic statements praising Adolf Hitler.
Companies like Google and OpenAI, which also use AI image generators, have many stricter rules around content.
The proliferation of inconsistent deepfake images has coincided with the widespread adoption of artificial intelligence. 400% increase According to the Internet Watch Foundation, in the first half of 2025, AI will use images of child sexual abuse.
xAI presented “Acute modein its image and video creator tool in August for verified adult subscribers to create sensual content.
According to Copyleaks, an artificial intelligence text and image detection company, some adult content creators on X encouraged Grok to create sexual images to promote themselves, which started an online trend a few days ago.
Testing Grok's capabilities went public as users asked it to create sexy images of celebrities and others.
xAI is reportedly valued at more than $200 billion and is investing billions of dollars to build the world's largest data center to support its artificial intelligence applications.
However, Grok's capabilities still behind competing AI models such as ChatGPT, Claude and Gemini, which have attracted more users, while Grok turned to AI sexual companions and risky chats to accelerate growth.
