App Store Review Guidelines

Apple revised its App Review Guidelines twice in 2026, and both updates matter if you're shipping an iOS app this year. The direction is consistent: stricter quality bars, more scrutiny on AI features, and less tolerance for apps that don't clearly add value.

The February 2026 update: user-generated content

In February, Apple clarified that apps offering random or anonymous chat features fall under guideline 1.2, the User-Generated Content section. This closes a gray area that some apps had been operating in — if your app connects strangers for open-ended conversation, you now need to meet the same moderation, reporting, and content-filtering requirements as any other UGC platform, even if chat wasn't the app's primary feature.

The June 2026 update: AI frameworks and export compliance

The bigger revision landed in June, updating both the Apple Developer Program License Agreement and the Review Guidelines. Key changes:

• Developer identity for export compliance — clarified requirements around disclosing developer identity information tied to export compliance rules.
• Sensitive Content Analysis framework — tightened requirements for apps using this framework, which flags explicit images on-device.
• Suggested Actions API — new specified requirements for apps that use this API to surface AI-driven suggestions to users.
• Trust Insights framework — new requirements for apps using this newer framework, part of Apple's broader push to formalize how apps handle trust and safety signals.

If your app touches any of these — and an increasing number of apps with AI features do, even indirectly — it's worth re-reading the relevant guideline sections before your next submission rather than assuming last year's compliance still holds.

The crackdown on "well-established" low-effort categories

Perhaps the most consequential change for indie developers and small studios: Apple named specific categories — dating apps, flashlight apps, sound effects apps, wallpaper apps, simple timers, and fortune-telling apps — as "well established," meaning new submissions in these categories will be rejected unless they offer a meaningfully different or improved experience over what already exists. Existing apps in these categories aren't safe either; Apple says it may remove them if they aren't updated, improved, or fail to attract customers.

A related line targets "mediocre, low-quality, or low-effort" apps by name, calling out things like drinking games and novelty apps specifically. The warning is blunt: repeated low-effort submissions can lead to removal from the Apple Developer Program entirely, not just rejection of the individual app.

The specific AI data-sharing disclosure rule catching developers off guard

A separate but closely related tightening deserves its own callout because it's tripping up developers who assume generic privacy language still covers them: Apple now requires apps to disclose, by name, which specific third-party AI provider receives user data — generic phrasing like "we may share data with service providers" is explicitly no longer sufficient. If an app sends user content to OpenAI, Google, Anthropic, or any other specific model provider for processing, the app's privacy disclosure needs to name that provider directly, and the app must obtain explicit user consent before that data sharing happens, not rely on a blanket privacy-policy acceptance from onboarding. This closes a gap a lot of apps with AI features have been quietly operating in — bolting a third-party LLM API onto a feature without updating the privacy nutrition label to reflect exactly where user data is actually going.

The compliance timeline is tight: developers are expected to have data privacy disclosures — including clear data usage policies, provider-level disclosure for any AI integration, and user-facing options for data deletion — fully implemented by Q3 2026 to stay ahead of tightening enforcement across major app stores. For any app using AI functionality, however incidental to the core product, this means an audit of every third-party API call that sends user data off-device is now a required pre-submission step, not an optional nicety — a chatbot feature, a smart-suggestion feature, or even background AI-assisted content moderation each need this disclosure if they route data to an external provider.

What actually gets apps rejected, by the numbers

Context on scale helps calibrate how much this all matters in practice: Apple reviewed over 9.1 million app submissions in 2025 and rejected roughly 2.09 million of them — close to a 23% rejection rate, continuing a year-over-year climb from 2024's roughly 25% rate on 7.77 million submissions. Ranked by frequency, the four biggest rejection categories are performance problems (crashes, bugs, incomplete builds, or dead demo accounts, under Guideline 2.1), legal and privacy issues (Guideline 5.1.1 — directly relevant to the AI data-sharing disclosure requirement above), design issues including minimum-functionality and spam/copycat detection (Guidelines 4.2 and 4.3, the section most likely to catch template-based or low-code apps), and business-rule violations around in-app purchase implementation (Guideline 3.1.1).

On timing: Apple states roughly 90% of submissions get reviewed within 24 hours, but real-world Q1 2026 data from developers shows new app submissions commonly taking 2-5 days, with spikes past a week during peak submission periods, while routine updates to already-approved apps typically clear in 24-72 hours. The rejections aren't usually permanent, either — 387,087 submissions were approved in 2025 after an initial rejection, meaning most rejection reasons are genuinely fixable rather than a dead end, provided the underlying issue (not just the symptom Apple flagged) actually gets addressed before resubmission.

What this means in practice

Apple's review process in 2026 is visibly oriented around three things: data transparency, AI feature disclosure, and genuine user value. Generic clones and template apps that used to slip through are facing much higher rejection rates. Before submitting:

• If you're entering a "well established" category, be ready to articulate — in your submission notes and in the app itself — what's genuinely different about your app.
• If your app uses on-device AI frameworks (Sensitive Content Analysis, Suggested Actions, Trust Insights), review the specific guideline language tied to each; generic AI feature descriptions are no longer sufficient.
• If your app has any chat, messaging, or anonymous connection feature, assume UGC guidelines apply and have moderation/reporting tooling in place before submission, not after.
• Budget extra review time. Stricter guidelines generally mean longer review cycles and more requests for clarification, especially for apps that touch AI functionality or open-ended user content.

The overall message from Apple this year is less about new restrictions and more about closing loopholes — apps that were technically compliant but low-value are no longer getting a pass just because they don't explicitly break a rule.

Sources: MacRumors — App Store Guidelines Low-Quality Apps, 9to5Mac — Apple Tightens App Review Guidelines, Apple Developer News, techrepublic.com, techcrunch.com, qawerk.com, appcompliance.io