(Credit: Thomas Fuller/SOPA Images/LightRocket via Getty Images))
Days after governments worldwide began scrutinizing X (formerly Twitter) for allowing its AI platform Grok to generate images of real people in bikinis without their consent, the platform announced it would prevent the chatbot from doing so altogether.
“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,” X’s Safety team said on Wednesday, adding that “this restriction applies to all users, including paid subscribers.”
Elon Musk's xAI startup will also block the standalone Grok app's ability to generate nudity “in those jurisdictions where it’s illegal.”
The issue stems from an X trend that led several users to ask Grok to generate bikini images of other users, including minors. The trend raised safety and privacy concerns, prompting governments in Indonesia and Malaysia to block the chatbot. A few other governments either launched an investigation or demanded stricter restrictions.
In the immediate aftermath, X decided to limit Grok’s image generation to paid users. It also vowed to ban accounts and take legal action against those involved in generating child sexual abuse material (CSAM).
The scrutiny, however, continued. This week, California’s attorney general, Rob Bonta, launched an investigation into the chatbot and noted that potential victims could file a complaint against xAI, which developed Grok. The investigation will also examine the standalone Grok app’s "Spicy Mode,” which enables users to generate images of AI avatars in minimal clothing.
Though not a formal response to the AG’s investigation, Musk said that with NSFW enabled, Grok is still allowed to generate upper-body nudity involving fictional adults, “consistent with what can be seen in R-rated movies on Apple TV.” He added that it is “the de facto standard in America.”
Earlier this week, the US Senate passed a bill allowing victims of nonconsensual deepfake imagery to sue its creators. President Trump has already signed a bill that requires social media platforms to remove such images within 48 hours of receiving notice.
Meanwhile, The Verge found that Grok was still able to generate revealing deepfakes on Wednesday. Musk, on the other hand, claimed that he was not “aware of any naked underage images generated by Grok. Literally zero.” He continues to blame users for making such requests and notes that “adversarial hacking of Grok prompts” may deliver “something unexpected.”