Apple updates App Store AI code-of-conduct and new opt-in labeling

AI · 3 min read

The revised App Store guidance requires apps that deliver generative outputs to disclose high-level model provenance and whether outputs may be inaccurate or biased. Developers must also provide an in-app mechanism for users to flag problematic content and a visible label on product pages indicating the presence of AI-generated features.

Apple's guidance emphasizes that apps must not misrepresent AI outputs as verified facts and must provide context such as confidence levels or source citations where applicable. For apps that use third-party models, developers are asked to summarize data handling and opt-out choices for user-contributed content used to fine-tune models.

Enforcement will begin with a phased rollout, giving developers time to update metadata and privacy notices. Apple is also expanding its review checklist to include an AI risk assessment and updated reviewer training for generative feature evaluation.