Grok’s creator, Elon Musk’s artificial intelligence company xAI, did not respond to WIRED’s request for comment about sex videos created with Grok Imagine. Since Grok began flooding X’s social media platform with AI-generated sexual images of women and people appearing to be underage more than a week ago, Musk and X said it was taking action against material depicting child sexual abuse. “Anyone who uses Grok to create illegal content will face the same consequences as those who upload illegal content.” Musk posted on X.
Like other tech companies that consistently battle the CSAM deluge, xAI’s policies state that “sexualization or exploitation of children” is prohibited on its services, as is “any illegal, harmful or abusive activity.” The company also has processes in place to detect and mitigate emerging CSAM materials. In September A Business Insider reportabout which the outlet claims to have interviewed 30 current and former xAI employees, determined that 12 of those employees “encountered” both sexually explicit content and written AI CSAM prompts on its services. Employees described systems that attempt to detect AI CSAM and prevent AI models from being trained on the data.
Apple and Google, which make Grok available in their app stores, did not respond to WIRED’s request for comment. Netflix also did not respond to a request for comment.
Unlike other major generative AI companies like OpenAI and Google, xAI enabled Grok to create AI-powered pornography and adult material. Previous reports have noted how you can create hardcore pornography with Grokwhich has a “spicy” mode. “If users select certain features or enter suggestive or vulgar language, the Service may respond with dialogue that may include vulgar language, vulgar humor, sexual situations or violence” – xAI Terms of Service to talk.
“Over the last few weeks and now we feel as if we have gone off a cliff and are free-falling into an abyss of human depravity,” says Clare McGlynn, a professor of law at the University of Durham and an expert on image-based sexual abuse, who says she is “deeply concerned” by Grok’s videos. “This technology supports and facilitates the inhumane impulses of some people, without protective barriers or ethical guidelines.”
McGlynn says allowing AI-generated pornography – which is not intended to depict a specific, real person – raises a number of questions about the safeguards put in place to prevent potentially illegal pornography, such as depictions of bestiality or rape, and the impact it could have. “What becomes an issue for me is the impact of creating and sharing pornographic content freely available to everyone, normalizing and minimizing sexual violence,” says McGlynn, while noting that explicit AI-powered images and videos of real people are already illegal in many countries.
Unlike X, which requires someone to log in if a post has been flagged as containing “age-restricted adult content,” Grok does not place any age restriction in order to display sexually explicit videos generated on the platform. Many states in the U.S. have recently passed age verification laws that require websites to verify the age of users if more than a certain percentage of the site’s content is sexually explicit.