Three Democratic senators urged Apple and Google to remove Elon Musk's apps X and Grok from their app stores Thursday evening after xAI's Grok artificial intelligence tool had been used to flood X with sexualized nonconsensual images of real people. Hours later, X adjusted how Grok operated on the social media site, restricting its image generation to paying premium subscribers, and seemingly restricting what types of images Grok can create on X.
The Grok reply bot on X has churned out thousands of sexualized images an hour this week, mostly of women but at times of children. Early Friday, it appears to have pivoted to limiting that feature on the social media app. But on the standalone Grok app and website, Grok will still create sexualized deepfakes.
In an open letter to Apple CEO Tim Cook and Google CEO Sundar Pichai, Sens. Ron Wyden of Oregon, Ed Markey of Massachusetts and Ben Ray Luján of New Mexico asked the companies to "enforce" terms of service that appear to ban the activity that was surging on X and is still possible on Grok.
The terms of service of Apple's App Store and Google's Play Store both appear to forbid apps that allow sexualized images of people without their consent, the senators wrote. "Apple and Google must remove these apps from the application stores until X's policy violations are addressed."
For more than a week, users have prompted the official Grok reply chatbot to generate sexualized images of nonconsenting people, putting them in more revealing clothing such as swimsuits and underwear. "X users have used the app's Grok AI tool to generate nonconsensual sexual imagery of real, private citizens at scale," the senators wrote. "This trend has included Grok modifying images to depict women being sexually abused, humiliated, hurt, and even killed."
"Turning a blind eye to X's egregious behavior would make a mockery of your moderation practices. Indeed, not taking action would undermine your claims in public and in court that your app stores offer a safer user experience than letting users download apps directly to their phones," they said.
Friday morning's move by X seems at least partly in response to sustained backlash against Grok's production of sexual deepfakes, but Musk and X have not indicated that there will be a wider rollback of Grok's capabilities on all platforms, including the downloadable Grok app, which remains in the Google and Apple app stores.
On Sunday, Musk and X reiterated that making illegal content will result in expulsion from the platform, though most of the content that was being made by the chatbot did not fit into that category.
Apple's terms of service for its App Store say that "Apps should not include content that is offensive, insensitive, upsetting, intended to disgust, in exceptionally poor taste, or just plain creepy." That includes "Overtly sexual or pornographic material, defined as 'explicit descriptions or displays of sexual organs or activities intended to stimulate erotic rather than aesthetic or emotional feelings.'" The App Store also says apps should not tolerate "defamatory, discriminatory, or mean-spirited content," particularly "if the app is likely to humiliate, intimidate, or harm a targeted individual or group."
Google's terms of service for its Play Store say that it does not "allow apps that contain or promote content associated with sexually predatory behavior, or distribute non-consensual sexual content."
Both Apple and Google have previously removed apps devoted to "nudifying" images of real people.
Grok and X are currently both popular on the Google and Apple app stores. Friday morning, Grok, where the sexual deepfakes are still seemingly allowed to be made, was ranked no. 4 in Apples App Store and No. 10 in Google's.
Neither company responded to a request for comment about the senators' letter, nor to previous NBC News questions about how the companies are considering X's role in nonconsensual sexualized imagery.
Musk, the owner of both X and xAI, the AI company that powers Grok, has long campaigned against heavy moderation, which he has equated with "censorship." In December, he unveiled a version of Grok that would manipulate images of real people.
An NBC News review of some of the deepfake images churned out this week found that most are of women who are depicted as wearing skimpy clothing, but some are of children. In some images, users successfully prompted Grok to put people in transparent or semi-transparent underwear, effectively making them nude.