Apple bans 'nudify' apps but leaves Grok in the App Store

Apple has updated its App Store Review Guidelines to explicitly prohibit apps that use AI to create intimate images of people without their consent, the so-called "nudify" apps. Its previous rules already banned apps designed to harass users. The new wording names this category directly. Developers can no longer argue that they are only offering "image editing" or "AI enhancement."

The change follows a year of mounting pressure. Earlier this year, X users found that Grok, the platform's chatbot, could generate sexual deepfakes of real people, including minors. According to NBC News, Apple told US senators it had threatened to remove Grok from the App Store. At WWDC in June, advocacy groups protested demanding that Grok, X and all nudify technology be pulled. In July, the San Francisco city attorney ordered Apple to remove 13 specific apps of this kind. Apple complied, as The Verge reported. That episode showed the existing guidelines were too vague for the new category.

The update is notable for what it leaves out. Grok and X remain available. Apple appears to separate apps whose main purpose is generating nonconsensual imagery from general platforms where that abuse is possible but not the core function. Today that line is defensible. Tomorrow it will be fragile, because image generation is being built into more and more mainstream apps. Banning apps that only undress people is easy. Policing a general-purpose assistant that can do the same thing when asked is much harder.

Enforcement is the second weak point. Apple's App Review has always been opaque. Deciding whether an AI app can be misused is not the same as checking whether it contains explicit content. The source also notes that the nudify ecosystem has proved resilient: when Apple and Google remove apps, developers rebrand them and submit them again. Critics argue Apple acted only once public pressure made doing nothing untenable. Its move can also be read as a way to get ahead of legislation by showing it can regulate itself.

Our reading: the measure is necessary and arrives late, but it points in the right direction. It confirms a pattern we have described before. The technical barrier to this kind of abuse has disappeared. The tools are within reach of any teenager, as recent school cases in Latin America have shown. So control has to move from the illegal file to the route that leads to it: who offers the capability, how it is distributed and how easy it is to use. Apple's closed ecosystem gives it unusual leverage over that route, and therefore greater responsibility than a filter after the fact.

The next step is to apply the same standard by capability, not by app label. A platform whose built-in assistant generates intimate deepfakes on request should face the same requirements as a nudify app, with effective safeguards, auditing and consequences. Generative imaging has enormous creative and productive value, and that value will only last if the harms to real people, especially minors, are dealt with now. Here, evidence-based governance means looking at what a tool actually does, not at how it markets itself.