Grok Still Hosting Sexualized Deepfakes of Famous Women Despite Earlier Safeguard Pledges

By Decode Today News

Elon Musk's AI chatbot, Grok, continues to host and facilitate the creation of non-consensual explicit images and videos featuring women, including public figures, months after its parent company xAI announced the implementation of restrictions to combat such harmful content. These revelations emerge as SpaceX, also owned by Musk, prepares for one of the largest IPOs in history this Friday, intensifying scrutiny on the entrepreneur's various ventures. A comprehensive analysis by WIRED of publicly available creations on Grok.com found numerous links leading to sexualized AI-generated images and videos. These depict celebrities and at least one politician in distressing scenarios, including being held against their will by a giant man, performing sex acts, and in states of full nudity. While some content appears animated, a significant portion is photorealistic, presenting scenarios that could plausibly be mistaken for real-world events. These findings suggest that despite xAI's stated commitments, the platform’s safeguards against the creation and hosting of non-consensual sexualized deepfakes remain insufficient. The problem extends beyond Grok.com, with some of these deeply concerning images and videos, generated via Grok Imagine, subsequently shared on X, Musk's social media platform. Although the exact timestamps of their creation are not visible on these posts, their recent appearance highlights a persistent issue. Experts suggest that the public examples reviewed by WIRED likely represent only a fraction of the content being generated, as creations are not made public by default. This raises questions about the scope of the problem and the effectiveness of current moderation. Industry experts are voicing serious concerns. Henry Ajder, a deepfake specialist with a decade of experience tracking explicit AI content, noted, "While Grok and X may have made some amendments to their model, particularly following the backlash around nudification at the beginning of the year, they still have not done a sufficient job to bring it up to the standard of the other mainstream tools that are available." This comparison underscores a significant gap in content moderation standards between Grok and its competitors.

A History of Controversy and Legal Challenges

xAI's Grok has been embroiled in controversy and faced legal challenges and regulatory scrutiny since January. At that time, its integration on X led to a surge of "nudification" images, where users, predominantly men, prompted the chatbot to alter images of women into "bikinis" or "string bikinis." A class-action lawsuit filed in California federal court in March even alleged that images of apparent minors were sexualized in some instances, bringing a critical child safety element into the debate. Following the initial outcry, xAI publicly committed to introducing safeguards to prevent the creation of non-consensual and sexualized deepfakes. The company has consistently reiterated its strict prohibition against Child Sexual Abuse Material (CSAM) on its platform. However, the latest findings by WIRED indicate that the issue of non-consensual adult deepfakes persists on Grok.com. When contacted for comment by WIRED regarding the recent findings, neither xAI nor X immediately responded. Notably, after WIRED's outreach, the explicit images and videos identified on Grok.com became unavailable, and the associated Grok Imagine links shared on X were removed, with X citing policy violations. In April, the safety account on X had stated, "We strictly prohibit users from generating nonconsensual explicit deepfakes and from using our tools to undress real people," in response to an NBC report on ongoing deepfake creation.

Prominent Targets and Alarming Prompts

Among the individuals depicted in the problematic content hosted on Grok.com, WIRED identified multiple celebrities and US Representative Alexandria Ocasio-Cortez. These videos portrayed the women in minimal clothing, held in the fist of a "giant" hand. One particularly disturbing prompt reviewed described a celebrity being held against her will, "pleading for him not to do this," as the "giant hand’s grip on her tightens... holding her in place as the giant man holding her leans in and licks her face up and down." To gauge the differing standards across AI platforms, WIRED tested two prompts used to generate material on Grok with other leading generative AI systems. OpenAI’s ChatGPT, Meta AI, and Anthropic’s Claude all rejected these prompts as inappropriate. Google’s Gemini, while creating an image for one celebrity held by a giant hand, also rejected another prompt. Google declined to comment on its system's response. Another concerning Grok Imagine video, also shared on X, appeared to depict Ashley St. Clair, who previously had a relationship with Elon Musk and is the mother of one of his children. The video allegedly showed her altered to be dancing in a bikini. St. Clair initiated legal action against xAI in January after similar sexualized deepfakes of her reportedly appeared on X. Following WIRED's contact with X, this post was also removed for violating platform rules. Carrie Goldberg, St. Clair's attorney and a prominent advocate against non-consensual sexualized imagery, highlighted the "historic" nature of this situation. While the creation of sexual abuse material featuring celebrities is not new, she noted that the ability to instantly generate and share such images on social media represents a critical escalation. "So there was instant publication of the digital nude images," Goldberg told WIRED. "And not only that, but the world’s richest man was promoting the nudification technology to his 240 million followers on X, monetizing it, and doing so on a platform that is for ages 13+." Imran Ahmed, CEO and founder of the Center for Countering Digital Hate (CCDH), echoed these concerns, claiming, "Elon Musk knowingly added a perverse feature to his platform that helps users undress women and children at the click of a button, with no regard for the predictable damage it would cause." In January, CCDH estimated that Grok had generated 3 million sexualized images, reportedly including over 20,000 of children. Ahmed added, "Now it appears explicit content is still being hosted on Grok and shared on X, including images ridiculing the mother of Musk’s child."

Lax Guardrails and Corporate Acknowledgments

Unlike many other major generative AI systems from companies like OpenAI and Google, Grok and xAI have maintained a more permissive stance on sexual content in general. Musk has previously stated that Grok is "supposed [to] allow upper body nudity of imaginary adult humans" and align with content seen in R-rated movies. The most recent terms of service from xAI acknowledge that the system "may respond with sexual situations." However, the company's documentation also explicitly states that its systems should not be used for "causing harm or engaging in abusive activity," creating a clear tension between its operational allowances and ethical guidelines. WIRED also observed other Grok Imagine videos, likely entirely AI-generated, depicting women undressing or engaged in sexual acts, some of which were entirely explicit. The user prompts for these videos often did not directly describe sexual acts but employed roundabout language, a likely tactic to bypass any existing safeguards on the Grok platform. Multiple researchers indicated to WIRED that since January, changes introduced by X and Grok appear to have made it more difficult to create "nudification" or "undress" images of real people. The number of such images posted to X has reportedly decreased in recent months, and users on Reddit and dedicated AI deepfake forums have noted increased moderation from the SpaceX-owned companies. Despite these reported changes, the ongoing challenges are substantial. In May, SpaceX, in a filing to potential investors, disclosed that it had set aside $530 million to address ongoing legal complaints, including those related to Grok. The filing cautioned, "Because these modes may be more irreverent and harsher than our standard offerings, they present heightened risks, including reputational harm, the generation of potentially explicit content and misinformation or deceptive outputs, potential nonconsensual or exploitative imagery, intellectual property infringement, or content that could be viewed as exploitative, harmful, harassing, abusive, or discriminatory."

The Road Ahead for AI Ethics

Adding to the pressure, ahead of SpaceX’s IPO this Friday, the Privacy Commissioner of Canada published preliminary findings from its investigation into xAI and Grok’s deepfake transgressions from January. The investigation alleged that xAI violated Canada’s federal private-sector privacy law by failing to implement "appropriate safeguards from the outset." While xAI informed the Privacy Commissioner that it has introduced new safeguards to prevent image alteration and "further proactive checks" for infringing content on social media, the Commissioner remains unconvinced. The investigation stated that "The respondents have not, to date, demonstrated the effectiveness of these safeguards in preventing and mitigating this issue." The continued presence of sexualized deepfakes on Grok underscores a critical challenge facing the rapidly evolving field of artificial intelligence: how to balance innovative technological capabilities with robust ethical guidelines and safety protocols. As generative AI becomes more sophisticated, the responsibility of companies to prevent misuse and protect individuals from harm grows exponentially. The incident highlights the urgent need for clear, effective, and transparent moderation policies that truly safeguard users and adhere to global ethical standards, rather than merely responding to public backlash. The ongoing scrutiny from regulators, legal challenges, and public advocacy groups signals that the spotlight on AI safety and accountability will only intensify.

Decode Today News

Ads

Breaking News

Grok Still Hosting Sexualized Deepfakes of Famous Women Despite Earlier Safeguard Pledges