Apple Tightens App Store Rules on Nonconsensual AI Nudify Apps

Apple has updated its App Store guidelines to explicitly ban apps that use AI to generate sexual images of people without their consent, according to The Verge.

The revised language says developers "must proactively prevent and remove" nonconsensual sexual images. The duty is on the developer to build prevention into the app, not only to respond to complaints. That means screening user prompts, checking AI outputs, and removing abusive images, whether the AI runs on the device or on company servers.

Failure can lead to removal. Apple says developers who lack adequate safeguards, or who do not fix a violation after notice, may be removed from the App Store. Either condition alone is enough.

Apple also clarified its rule against bait-and-switch apps. The rule has long barred apps that behave differently after review than during it. It now covers apps that add or change features, including "nudification," after approval. This covers cases where a benign app passes review and the harmful capability is added later through a server-side update, remote configuration, or model swap.

The update also addresses exploitation more broadly. Apple says apps that facilitate human trafficking or the exploitation of children will be removed from the App Store and blocked from alternative distribution. That second part extends enforcement beyond the store itself to other installation paths Apple controls.

The change follows months of pressure and reporting. In April, Apple and Google continued to offer apps that let users make nonconsensual sexualized images despite policies against them, according to Bloomberg. In June, protesters at Apple's WWDC event called on the company to remove Grok, X, and all nudify technology from the App Store. In July, San Francisco's city attorney ordered Apple to remove 13 nudify apps, and Apple said it removed them.

That July action was part of a wider demand. The San Francisco City Attorney's Office sent cease-and-desist letters to Apple and Google telling them to stop profiting from 13 face-swap apps, according to Wired. Apple did not remove X or Grok from the App Store after X users used Grok to create nonconsensual sexualized deepfakes.

The broader context here is enforcement, not wording. App review is a check at one point in time. Generative features are live systems. A prompt filter, safety classifier, or model version can change without a new app version, which makes monitoring after approval hard for any store operator. Clearer language gives Apple a stronger basis for takedowns, but it does not solve detection at scale.

In my view, the practical effect is to make safety work a compliance requirement for developers. Logging, abuse reporting, rapid removal, and version control for models and prompts move from best practice to baseline. That should reduce uncertainty for legitimate makers of image-editing, dating, and social apps, while raising costs for operators that rely on evading review.

Looking beyond the store, store policy alone has limits. Alternative distribution, web apps, and sideloaded models sit outside the same review pipeline. Blocking alternative distribution in trafficking or child exploitation cases narrows that gap for the most serious harms. For other nonconsensual imagery, prevention will still need cooperation across model providers, app developers, and store operators. The language is now plainer. Enforcement is the test.