SpaceX IPO Faces Scrutiny as Grok Deepfake Allegations Resurface
As SpaceX prepares for its initial public offering, new evidence suggests Elon Musk’s Grok chatbot continues to host nonconsensual sexualised imagery, drawing criticism from experts and regulators who argue safeguards remain insufficient.
A WIRED investigation has identified dozens of nonconsensual, sexualised deepfake images and videos hosted on Grok.com, the website for Elon Musk’s xAI chatbot. The findings emerged shortly before SpaceX’s initial public offering on Friday, adding a layer of regulatory and reputational risk to one of the largest market debuts in recent history. The content includes photorealistic depictions of celebrities and US Representative Alexandria Ocasio-Cortez, as well as AI-generated figures, raising questions about the efficacy of safety measures introduced earlier this year.
The investigation reviewed hundreds of public Grok Imagine links, finding that several led to sexualised AI content, including videos depicting women being held by a "giant" man. Specific content identified included altered images of Ashley St. Clair, the mother of one of Musk’s children, depicting her dancing in a bikini. Some links were shared on the social media platform X, though posts were removed after WIRED contacted the companies. The Centre for Countering Digital Hate estimated in January that Grok had created approximately 3 million sexualised images, including over 20,000 allegedly involving children.
US privacy regulators in Canada have published preliminary findings alleging xAI violated federal privacy laws by failing to implement appropriate safeguards from the outset. The investigation, released by the Privacy Commissioner of Canada, states that xAI has not demonstrated the effectiveness of its new safeguards in preventing and mitigating the issue. This regulatory scrutiny coincides with SpaceX’s IPO filing, which set aside $530 million to handle ongoing legal complaints, including those linked to Grok.
Experts note that while "nudification" of real people appears to have decreased since January, Grok still allows general sexual content and has fewer safety guardrails than competitors like OpenAI or Meta. Henry Ajder, an expert on deepfakes, stated that while Grok and X have made some amendments, they have not done a sufficient job to bring the platform up to the standard of other mainstream tools. Musk has previously stated that Grok is intended to allow upper-body nudity of imaginary adult humans and be consistent with R-rated movie content.
The controversy follows significant backlash in January when Grok was used to create a flood of "nudification" images, including alleged depictions of minors. xAI stated it had introduced safeguards to limit nonconsensual and sexualised deepfakes, banning child sexual abuse material. However, the Center for Countering Digital Hate CEO Imran Ahmed claimed Musk knowingly added features that help users undress women and children, with no regard for the predictable damage. As SpaceX goes public, the company faces heightened risks including reputational harm and potential nonconsensual or exploitative imagery, as outlined in its May filing.
