Apple's New App Review Rule 2.5.19 Bans Hidden and Dormant Features: What iOS Teams Must Change

YuSMP Group Editorial Team Editorial team, YuSMP Group

The short version

What changed in the App Review Guidelines?

Apple published three changes on October 16. The headline addition is guideline 2.5.19, which reads that apps must not include "hidden, dormant, or undocumented features" and that an app's functionality "should be clear to end users and App Review." Attempting to hide undocumented features in the binary, the rule continues, "such as through code obfuscation or other methods," is grounds for removing the app and terminating the developer account.

Apple frames the rule as a clarification of a long-standing ban on apps that look harmless during review and change after approval, a tactic it calls bait-and-switch. The second change extends guideline 1.2 on user-generated content: it now names nonconsensual intimate imagery and explicitly covers AI content creation apps. The third updates Section 5 so that apps found to facilitate human trafficking or child exploitation are removed from the App Store and blocked from alternative distribution, the route Apple runs for apps distributed outside its store in the EU.

The Verge, which reported the update the same day, highlighted the direct target: so-called "nudify" apps that pass review as photo editors and unlock sexualized deepfake features later.

Does guideline 2.5.19 affect feature flags and remote config?

Not as a technique, but it raises the bar on how they are used. Feature flags, staged rollouts, A/B tests and server-driven UI are standard practice in mobile delivery, and Apple's text does not mention them. What 2.5.19 forbids is functionality that App Review cannot see: code paths that stay dark during review and are activated afterwards, or capabilities deliberately disguised in the binary.

The practical line is materiality. Toggling a reviewed checkout flow for 10% of users is a rollout. Turning on a feature that the reviewed build never showed, such as a new payment method, a chat module or a content generator, is a change in what the app does. The second case should go through review as a new build, with the behavior described in the App Review notes and demo credentials that reach it. Obfuscation used for intellectual property protection is still common; the risk starts when it hides behavior rather than implementation.

What do AI image and user-generated content apps have to do?

Guideline 1.2 now applies to apps that let users "create, generate, upload, share, or otherwise contribute" content, and lists nonconsensual intimate imagery next to IP infringement and bullying. Apple's announcement says such apps must proactively prevent and remove this content, including imagery generated with AI tools. The existing requirements stay: filtering of objectionable material, a reporting mechanism with timely responses and the ability to block abusive users. Developers who lack adequate safeguards, or do not remedy violations after notice, risk removal under the Developer Code of Conduct.

For any app built around image generation or editing, that means moderation can no longer be a post-launch backlog item. Prompt filtering, output classification and a working report-and-takedown loop need to exist in the build that App Review sees.

Why is Apple tightening review now?

Apple ties the update to fraud that only activates after approval. In its announcement it says that last year it prevented more than 150,000 apps with safety issues from reaching customers and terminated more than 193,000 developer accounts for fraud and abuse, and that some of these cases involved coordinated operations hiding harmful content until after review. Generative AI has made that pattern cheaper: a harmless-looking photo app can fetch a different model or prompt pipeline from a server at any time. Writing the ban into its own numbered rule gives Apple a clear basis for account-level enforcement.

What it means for US & EU software teams

For most product teams the rule changes paperwork more than architecture. The exposure sits in release practices that grew without review in mind: "kill switches" that also unlock unreleased features, experiments that ship a whole new module behind a flag, or white-label builds where client-specific functionality is enabled by configuration after approval. Each of these should be documented, or shipped as a reviewed build.

The account-level penalty changes the risk calculation for companies that publish several apps under one developer account. A single 2.5.19 violation in one product can, per Apple's wording, put the entire account at risk, which argues for a release owner who checks review notes against what the flags actually control.

EU teams have an extra angle. Apps distributed through alternative marketplaces still pass Apple's notarization review, and the updated Section 5 explicitly extends blocking to alternative distribution. Teams building AI image features for EU users also need to line up the new 1.2 obligations with their GDPR handling of personal images and their content-moderation duties under EU platform rules.

Checklist before your next App Store submission

1. Inventory your flags. List every remote-config key and feature flag and mark which ones enable functionality absent from the reviewed build.
2. Write honest review notes. Describe server-controlled behavior, staged features and geo-specific functionality, and give App Review a demo account that reaches them.
3. Ship new capability as a new build. If a flag would materially change what the app does, submit an update instead of flipping it in production.
4. Check obfuscation scope. Make sure obfuscation protects implementation details, not undisclosed features.
5. Harden UGC and AI moderation. Generative apps need prompt and output filtering, reporting, blocking and a documented takedown process in the submitted build.
6. Assign a release owner. One person signs off that the build, the flags and the review notes describe the same app.

Frequently asked questions

Sources

Apple Developer — Strengthening App Review Guidelines against nonconsensual intimate imagery and bait-and-switch apps (October 16, 2026)
Apple Developer — App Review Guidelines (updated October 16, 2026)
The Verge — Apple’s updated App Store rules go after ‘nudify’ AI apps (October 16, 2026)