Grok’s sexual deepfakes almost got it banned from Apple’s App Store. Almost.

Apple quietly threatened to kick Elon Musk’s AI app, Grok, from its App Store in January over its failure to curb the surge of nonconsensual sexual deepfakes flooding X, according to NBC News. It was a muted show of force from one of tech’s most powerful gatekeepers, made behind closed doors even as the undressing crisis unfolded in full public view and criticism over Apple’s cowardice mounted.

In a letter obtained by NBC News, Apple told US senators it “contacted the teams behind both X and Grok after it received complaints and saw news coverage of the scandal” and demanded that the developers “create a plan to improve content moderation.” At the time, xAI’s chatbot Grok was freely accessible on X and as a standalone app, with flimsy safeguards that allowed users to easily generate and share sexualized deepfakes and “undress” images of real people, disproportionately women and some of them apparently minors.

As we reported at the time, these were flagrant and unambiguous violations of App Store guidelines it often applies with an iron fist. Apple, which profits from having apps like X and Grok on its digital store, has not spoken publicly about the issue or its behind-the-scenes intervention. Google, through its Google Play app store, profits similarly and has also not commented publicly on the matter.

Apple said it reviewed proposed changes to the X and Grok apps. While the company concluded X had “substantially resolved its violations,” Grok “remained out of compliance.” Apple said it warned the developer that “additional changes to remedy the violation would be required, or the app could be removed from the App Store.” Only after further back and forth did Apple determine Grok had “substantially improved” and approved its submission.

Throughout this covert back-and-forth, Grok and X appear to have remained live on the App Store, a drawn out process that may help explain the confusing, haphazard rollout of moderation changes announced in real time. This included limiting Grok on X to paying subscribers and attempting to stop Grok from undressing women. Our investigations revealed that neither were particularly effective beyond making the tool a bit harder to access. Later interventions, like X letting users block Grok from editing their photos, are also easily circumvented.

Despite Apple’s approval and xAI’s claims it has tightened safeguards, Grok still appears to be able to generate sexualized deepfakes with relative ease. Cybersecurity sources told me they have been able to create explicit images of celebrities and political figures using the tool, and I have been able to produce similar images of myself and other consenting adults. NBC also reported similar findings yesterday.