Apple’s guideline 5.1.2(i) requires that any app sharing personal data with a third party, including a third-party AI provider such as OpenAI, Anthropic, or Google, must clearly disclose which provider and what data before doing so, and let the user revoke that permission later. It’s been enforced since 13 November 2025, so it isn’t new — but nearly a year in, it’s still a recurring cause of App Store rejections, because it targets exactly the pattern most common in apps that add an AI feature after the fact: a chat box or AI-generated content feature wired up to an external API without a corresponding consent flow.
Why this keeps catching teams out
The guideline doesn’t ban sending user data to AI providers — it requires the disclosure to be specific (naming the provider, not just “we use AI”) and the consent to be obtained before the data leaves the device, not buried in a general privacy policy accepted at signup. Teams that treat an AI integration as a backend implementation detail, rather than a data-sharing event with its own UX, are the ones most likely to build the feature, ship it, and get rejected at review — a costly place to discover a compliance gap.
So what
If a build includes any feature that sends user data to a third-party AI model — summarisation, chat, image generation, recommendations — the consent flow for guideline 5.1.2(i) needs to be designed alongside the feature, not added after a rejection. It’s a small piece of UX with real launch-timeline consequences if it’s missed. We scope AI-integrated features with App Store compliance built in from the start rather than treated as a review-stage fix — see our AI products work, or get in touch to talk through a build that includes AI-powered features.