Grok, a chatbot from Elon Musk's artificial intelligence company xAI, published sexual images of children after a lewd request from a user apparently caused its guardrails to fail.
Users used prompts like “change her into a bikini” under photos of real people on X to have Grok generate non-consensual images of them wearing inappropriate clothing. The transformed images created on Grok's account are published on Musk's social media platform, X.
The AI complied with requests to morph images of minors, even though that itself was a violation. terms of service.
“There are isolated cases, like the example you mentioned, where users requested and received AI images depicting minimally dressed minors,” Grok said. answered “xAI has safeguards in place, but improvements are underway to completely block such requests.”
xAI did not immediately respond to a request for comment.
The chatbot posted: apology.
“We deeply regret the incident on December 28, 2025, when, based on a user's prompt, we generated and shared an AI image of two girls (estimated ages 12-16) dressed in sexual attire,” Grok's profile states. “This violated ethical standards and potentially violated US law regarding CSAM. It was a failure of safety measures and we apologize for the harm caused. xAI is reviewing this to prevent future issues.”
The Indian government has notified Company X that it must submit a report within 72 hours on the steps it has taken to stop the production and distribution of non-consensual and indecent images of women or risk losing its legal immunity.
Critics have accused xAI of allowing AI-based harassment, and are shocked and angered by the existence of features for seamless AI manipulation and requests to disrobe.
“Why isn't this illegal?” journalist Samantha Smith posted on X, accusing her of creating non-consensual sexual photos of herself.
Musk's xAI positions Grok as an “anti-wake” chatbot, programmed to be more open and edgy than competing chatbots such as ChatGPT.
In May, Grok posted about “white genocide” in response to an unrelated question, repeating the conspiracy theory that black South Africans are persecuting the white minority.
In June, the company apologized after Mr. Grok posted a series of anti-Semitic comments praising Adolf Hitler.
Companies like Google and OpenAI also operate AI image generators and have achieved a lot of success. More restrictive guidelines around the content.
The proliferation of non-consensual deepfake images has coincided with the widespread adoption of AI. 400% increase According to the Internet Watch Foundation, AI-generated images of child sexual abuse increased in the first half of 2025.
Introduction of xAIspicy mode” was added to our image and video generation tool for creating erotic content for verified adult subscribers in August.
The internet trend started a few days ago when some adult content creators on X encouraged Grok to generate sexual images to promote themselves, according to AI text and image detection company Copyleaks.
Testing Grok's limits evolved into a free test as users asked to create sexually explicit images of celebrities and others.
xAI is reportedly valued at more than $200 billion and is investing billions to build the world's largest data center to power its AI applications.
However, Grok's capabilities still late Competing AI models such as ChatGPT, Claude, and Gemini are gaining more users, but Grok turned into a sexual AI companion A racy chat to drive growth.
