Elon Musk's Grok chatbot continues to host nonconsensual explicit images and videos of women, months after his AI company xAI said it would introduce restrictions to stop the creation of sexualized deepfakes. The revelations come as SpaceX, xAI's parent company, prepares to go public on Friday in one of the largest IPOs of all time.
A WIRED analysis of public creations found that the Grok Imagine generative AI system has been used to create and host images and videos depicting celebrities and at least one politician. Some content shows women being held against their will by a giant man, performing sex acts, or appearing fully nude. While some images are fully AI-generated or in animated styles, others are photorealistic and depict plausible real-world scenarios.
WIRED reviewed hundreds of public Grok Imagine links hosted on Grok.com and found dozens led to sexualized AI images and videos, including those created without the subject's consent. Some links were subsequently shared on X in recent days. The posts are likely just a snapshot of what is being created, as generations do not appear to be made public by default.
Guards Still Lax Compared to Rivals
Other generative AI systems deploy more safety guardrails than Grok, which is available on its own website and on X. Henry Ajder, an expert on deepfakes who has tracked explicit AI content online for nearly a decade, said that while Grok and X may have made some changes after backlash earlier this year, they have not done enough to match the standards of other mainstream tools.
Musk's xAI has faced a wave of lawsuits and scrutiny from regulators around the world since January, when the Grok implementation on X was used to create a flood of nudification images. People on X, primarily men, asked the chatbot to edit images to show women in bikinis. In some cases, images of apparent minors were allegedly also sexualized, according to a class-action lawsuit filed in California federal court in March.
Since then, xAI has said it has introduced safeguards to limit and prevent the creation of nonconsensual and sexualized deepfakes. The company has consistently said child sexual abuse material is banned on its platform. Neither xAI nor X immediately responded to WIRED's request for comment. After WIRED contacted the companies, the explicit images and videos hosted on Grok.com appeared to be no longer available, and links shared on X were removed for policy violations.
Celebrities and Politician Targeted
Among the posts hosted on Grok.com, WIRED identified images and videos depicting multiple celebrities as well as US Representative Alexandria Ocasio-Cortez. Spokespeople for Ocasio-Cortez did not immediately respond to WIRED's request for comment. In the videos, the women were depicted wearing little clothing while being held in the fist of a giant man. One prompt described a celebrity being held against her will as she pleads for him not to do this, with the giant hand tightening as the giant man leans in and licks her face.
Two prompts used to generate material on Grok were rejected by OpenAI's ChatGPT, Meta AI, and Anthropic's Claude as inappropriate when tested by WIRED. Google's Gemini created an image of one celebrity being held in the hand of a giant but rejected another prompt. Google declined to comment.
Stay updated
Get the day's AI and automation news in your inbox. No spam, unsubscribe anytime.
One Grok Imagine video, also posted to X, appeared to depict Ashley St. Clair altered to be dancing in a bikini. St. Clair was previously in a relationship with Musk and is mother to one of his children. In January, she started legal action against xAI after sexualized deepfakes of her allegedly appeared on X. After WIRED contacted X, the post was removed for violating its rules. Legal representatives for St. Clair did not immediately respond to a request for comment.
Imran Ahmed, CEO and founder of the Center for Countering Digital Hate, said Musk knowingly added a perverse feature that helps users undress women and children at the click of a button. The CCDH estimated in January with high confidence that Grok created 3 million sexualized images, allegedly including more than 20,000 of children.
SpaceX Financial Risks Mount
Unlike other generative AI systems, Musk's Grok and xAI have not backed away from allowing sexual content in general, having previously introduced Spicy and Unhinged modes. Musk has stated Grok is supposed to allow upper body nudity of imaginary adult humans and be consistent with what viewers might see in R-rated movies. xAI's documentation says the system may respond with sexual situations but does not allow people to use its systems for causing harm or engaging in abusive activity.
Other Grok Imagine videos seen by WIRED show women, likely entirely AI generated, undressing or involved in explicit sexual acts. User prompts sometimes describe them in roundabout ways to circumvent safeguards.
Multiple researchers told WIRED that since January, changes introduced by X and Grok have made it harder to create nudification images of real people. The number of these images posted to X appears to have decreased in recent months. On Reddit and a dedicated AI deepfake forum, users complained about increased moderation.
Nevertheless, in May, SpaceX warned potential investors that it has set aside $530 million to handle ongoing legal complaints, including those linked to Grok. The filing said these modes present heightened risks including reputational harm, generation of potentially explicit content, misinformation, nonconsensual imagery, intellectual property infringement, and content that could be viewed as exploitative or discriminatory.
Ahead of SpaceX's IPO on Friday, the Privacy Commissioner of Canada published preliminary findings of an investigation into xAI and Grok's deepfake transgressions in January. The investigation alleged xAI violated Canada's federal private-sector privacy law by not including appropriate safeguards from the outset. The investigation says xAI told the Privacy Commissioner it has introduced new safeguards to prevent people from successfully prompting Grok to alter clothing in images and has introduced further proactive checks of social media websites to find infringing content. However, the Privacy Commissioner is not convinced by the changes, saying the respondents have not demonstrated the effectiveness of these safeguards.
Related on Neura Market
- AI Models Directory, Explore other AI models and their safety features.
- AI Image Tools, Compare generative AI tools and their content policies.
- Deepfake Detection Resources, Find tools and information on identifying synthetic media.