Apple review forces AI feature cuts

The fight here is about iPhone apps, but really it’s about who gets to decide what “safe AI” looks like. Apple runs a curated store, and that means every new AI feature has to survive human review before it reaches users. The gap is that AI apps move fast, while App Review has always been slower, fuzzier, and built around broad rules rather than product-specific playbooks. (forbes.com) Now developers say that mismatch is getting worse as Apple tightens privacy and content enforcement around AI. (developer.apple.com) ### Why are AI apps hitting this wall? Because AI features trip several Apple rules at once. A chatbot can look like user-generated content. A smart assistant can send personal data to an outside model. An app that spins up lots of near-identical tools can start to look like spam. (forbes.com) Apple’s guidelines are broad on purpose, but that also means reviewers have wide discretion when a product sits in the gray zone. ### What changed on Apple’s side? The clearest shift came on November 13, 2025. Apple updated its App Review Guidelines to say apps must clearly disclose when personal data is shared with third parties, including third-party AI, and must get explicit permission first. Then on February 6, 2026, Apple added another clarification saying random or anonymous chat apps fall under its user-generated-content rules. (developer.apple.com) That combination makes a lot more AI products review-sensitive. ### Why does that feel bigger than one privacy rule? Because many AI apps are not just “using AI” in the abstract. They send prompts, screenshots, messages, documents, or profile data to outside model providers. Once Apple names third-party AI directly in the rules, review stops being a generic privacy check and becomes a feature-level compliance check. (developer.apple.com) Developers now have to explain data flow, consent, moderation, and edge cases much more concretely. ### Is this just developers complaining? Not entirely. There is a real supply shock underneath the anecdotes. Appfigures data cited by Forbes shows 557,000 new apps were submitted to Apple’s App Store in 2025, up 24% from 2024 and the highest annual volume since 2016. Apple says it still processes 90% of submissions within 48 hours on average and has handled more than 200,000 submissions a week over a recent 12-week stretch, with average review time around 1.5 days. (developer.apple.com) Both things can be true at once — the average can stay low while edge-case apps get stuck. ### Why would AI apps get stuck more than others? Because they are harder to classify. A calculator app is a calculator app. An AI app might be a writing tool, a social product, a search layer, a roleplay bot, a coding assistant, and a content generator all at once. (developer.apple.com) Reviewers then have to ask whether it needs moderation tools, clearer age gating, stronger disclosures, or proof that it isn’t just another clone. The catch is that each answer can force product changes, not just paperwork. ### Why are small developers feeling this most? Because they have less room for back-and-forth. A big company can afford compliance work, legal review, and multiple submission cycles. An indie builder might have one launch window and one version that has to work. If review asks for narrower permissions, more moderation, or fewer risky features, the fastest path to approval is often to cut the feature. (forbes.com) That turns App Review into product design by attrition. The complaints making the rounds fit that pattern. ### Does Apple have a reason to be stricter? Yes — and it’s not hard to see why. Apple is trying to keep the App Store from turning into a swamp of spammy, unsafe, or privacy-leaky AI software while it also pushes its own AI roadmap. (developer.apple.com) The problem is that broad safety rules can hit junk and legitimate experimentation at the same time. When the rules are clear only in hindsight, developers start designing for reviewability instead of usefulness. ### Bottom line? Apple has not announced a blanket anti-AI crackdown. But its late-2025 and early-2026 rule changes, plus a flood of new submissions, have made AI app review more fragile and more expensive. Basically, if your iPhone app depends on outside models, open-ended chat, or messy real-world user input, App Review is no longer the last step. (forbes.com) It’s part of the product. (developer.apple.com 1) (developer.apple.com 2)