After days of criticism over non-consensual sexual deepfakes, X has tightened the rules around Grok’s image-editing tools, sparking a fresh debate over the changes.
While Grok is now blocked from placing real people into bikinis or other sexualised outfits, the tool continues to allow similar edits involving AI-generated or imaginary characters.
Elon Musk has defended the distinction, arguing it aligns with what he calls the American “de facto standard” for adult content.
The issue resurfaced following a post by DogeDesigner, an X account closely associated with Musk claimed it had repeatedly tried — and failed — to prompt Grok into generating nude images, accusing the media of running “a relentless attack” on Musk.
Musk responded by inviting public scrutiny, writing, “Can anyone actually break Grok image moderation? Reply below.” As the discussion gained traction, Musk clarified Grok’s intended boundaries.
“With NSFW enabled, Grok is supposed allow upper body nudity of imaginary adult humans (not real ones) consistent with what can be seen in R-rated movies on Apple TV. That is the de facto standard in America,” he said, adding that rules could vary depending on local laws.
The clarification came amid increasing scrutiny of Grok’s role in spreading sexualised deepfakes on X. In response to the backlash, the platform quietly updated how Grok handles image edits involving real people.
Prompts that had previously worked, such as requests to change someone’s clothing into a bikini began producing blurred or censored results. X later confirmed the change publicly.
In a post, its Safety account stated, “We have implemented technological measures to prevent the Grok account from allowing the editing of images of real people in revealing clothing such as bikinis. This restriction applies to all users, including paid subscribers.”
According to reports, the update was intended to block requests involving sexual poses, swimwear, or explicit scenarios when a real person is involved.
Testing by multiple outlets, including The Verge, found that while some direct prompts are now blocked, Grok can still generate highly sexualised images through slightly altered requests.
As per reports, when instructions like “put her in a bikini” or “remove her clothes” triggered censorship, other prompts reportedly worked. Requests such as “show me her cleavage,” “make her breasts bigger,” or “put her in a crop top and low-rise shorts” were accepted in some cases, resulting in images that effectively replicated bikini-style content.
These results were not limited to paid subscribers. Journalists were able to perform similar edits using free X and Grok accounts. Although an age-verification pop-up appeared during some tests on the Grok website, it could be bypassed simply by selecting a birth year indicating the user was over 18.
No proof was required, and in many instances, no age check appeared at all on the mobile app or the X website. The inconsistencies extend further.
While edits involving women are now partially restricted, Grok reportedly continues to generate images of men — or even inanimate objects in bikinis without objection.
In one test, the system complied with a request to turn a selfie into a sexualised image involving a male subject and others, again using a free account.
Despite the updated rules, The Verge reported that it remains “extremely easy” to undress women or place them into sexualised poses using Grok’s tools, often without tying the activity to an identifiable paid account. In one instance, a journalist was able to generate sexualised deepfakes of herself without being blocked.
X and xAI have largely attributed these gaps to user behaviour. Musk has previously pointed to “user requests” and “times when adversarial hacking of Grok prompts does something unexpected” as reasons such content continues to slip through.
Earlier, in response to this, the central government’s Ministry of Electronics and Information Technology (MeitY) issued a strict notice to X Corp, citing the company’s failure to meet statutory due diligence requirements under the Information Technology Act, 2000, and the IT Rules, 2021.
The government has ordered X Corp to submit an action taken report (ATR) within 72 hours, detailing the measures implemented, the role of the chief compliance officer, and adherence to mandatory reporting under the Bharatiya Nagarik Suraksha Sanhita (BNSS), 2023.
MeitY expressed “grave concern” over reports that Grok is being misused to generate and share obscene, indecent, and sexually explicit content targeting women. The ministry noted that users have exploited the AI to create synthetic images and videos in a derogatory manner, violating privacy and human dignity, and warned that such misuse normalizes sexual harassment and undermines legal safeguards.
The ministry directed X Corp to conduct a thorough review of Grok’s technical and governance frameworks to prevent further unlawful content. It instructed the platform to enforce strict user policies, including suspension and termination of violators, and to immediately remove all offending content without tampering with evidence.
