Elon Musk’s Grok chatbot continues to be used to produce and host nonconsensual explicit images and videos of women, months after xAI said it would introduce restrictions to stop the creation of harmful sexualized deepfakes, according to a WIRED analysis. The revelations come as SpaceX, xAI’s parent company, prepares to go public in one of the largest IPOs of all time. For enterprise technology leaders procuring AI systems, the case underscores the importance of rigorous content moderation and legal risk assessment.
The Persistent Content Moderation Failure
WIRED reviewed hundreds of public Grok Imagine links on Grok.com and found dozens leading to sexualized AI images and videos, including those created without consent. The content depicted celebrities and US Representative Alexandria Ocasio-Cortez in scenarios ranging from scantily clad women held in a giant man's fist to explicit sexual acts. Some were photorealistic. Two prompts that generated material on Grok were rejected by OpenAI’s ChatGPT, Meta AI, and Anthropic’s Claude when tested by WIRED. Google’s Gemini created one image but rejected another. Henry Ajder, a deepfake expert, told WIRED: “While Grok and X may have made some amendments… they still have not done a sufficient job to bring it up to the standard of the other mainstream tools.”
Enterprise AI Risks and Legal Exposure
For enterprises considering AI tools like Grok for internal or customer-facing use, the incident highlights serious risks: reputational harm, generation of explicit content, misinformation, and intellectual property infringement. In May, SpaceX warned investors it has set aside $530 million to handle ongoing legal complaints, including those linked to Grok. The filing noted that Grok's “irreverent and harsher” modes present “heightened risks, including … potential nonconsensual or exploitative imagery.” Imran Ahmed, CEO of the Center for Countering Digital Hate, which estimated that Grok created 3 million sexualized images (including more than 20,000 of children), said: “Elon Musk knowingly added a perverse feature to his platform that helps users undress women and children at the click of a button.”
| Key Risk Metrics (from source) | Value |
|---|---|
| Legal reserves (SpaceX) | $530 million |
| Estimated sexualized images (CCDH) | 3 million |
| Estimated images of children (CCDH) | >20,000 |
| Class-action lawsuits filed | 1 (California, March) |
Regulatory Scrutiny and Inadequate Safeguards
The Privacy Commissioner of Canada published preliminary findings of an investigation into xAI and Grok, alleging xAI violated Canada’s federal private-sector privacy law by not including appropriate safeguards from the outset. xAI told the Commissioner it introduced new safeguards, but the investigation stated: “The respondents have not, to date, demonstrated the effectiveness of these safeguards.” Despite xAI’s claims that child sexual abuse material is banned, WIRED found explicit content still hosted on Grok.com weeks later. Posts on Reddit and deepfake forums indicate users have complained about increased moderation, but the problem persists.
Implications for Enterprise AI Procurement
For CTOs and digital transformation leaders, the Grok saga offers a cautionary tale: even well-funded AI systems can fail at content moderation if governance is not embedded from the start. Unlike Grok, other generative AI systems from OpenAI, Google, and Anthropic deploy more safety guardrails. When evaluating AI vendors, enterprises should demand evidence of robust safety testing, continuous monitoring, and clear liability frameworks. The $530 million legal reserve set by SpaceX is a stark reminder of the financial consequences of inadequate AI governance. As the IPO approaches, the incident may influence investor confidence and regulatory scrutiny of AI-powered products across industries.
Sources:
