Apple is launching new AI-powered features for its latest smartwatches that can listen to conversations all day, remember what was said, and automatically create summaries. These features, known as Siri Recap and Live Rewind, are arriving with the first developer beta of watchOS 27.3 for the Apple Watch Series 12 and Watch Ultra 4. They are part of the Audio Intelligence system announced in September, joining existing Sound Recognition and Shazam-based Music Recognition tools.
Siri Recap is described as the more controversial of the two features because it automatically takes notes from conversations throughout the day. When enabled, it listens to surrounding audio to generate a title, summary, and key points after each interaction, which users can view in the Recaps tab of the Siri app. However, the feature does not produce full transcripts or identify individual speakers. Apple states that Siri Recap is designed to exclude sensitive information like financial details and government-issued identification numbers. These summaries are automatically deleted after seven days unless a user chooses to save them, though users can export them to the Journal or Notes apps.
Users have the ability to control when Siri Recap operates by scheduling it based on time and location, such as during office hours, or by toggling it via the Control Center. Both Siri Recap and Live Rewind are optional and remain off until activated by the user.
Live Rewind allows users to review the last 15 seconds of a surrounding conversation. By double-pressing the Digital Crown, a text snippet appears showing what was recently said, which could help a user remember a waiter's specials or missed directions. Users can ask Siri questions about this text or save it. The snippet disappears about 30 seconds after the watch display dims if no action is taken. Activating Live Rewind produces an audible chime even in silent mode or with headphones, and it displays a full-screen animation and a microphone indicator to alert nearby people.
Regarding privacy, Apple says the system does not create or store conventional audio recordings. Raw audio is processed within a hardware-isolated Secure Exclave on the S11 chip and the paired iPhone. Apple claims that the operating system, apps, users, and Apple itself cannot access this raw audio. While Live Rewind processes information directly on the device, Siri Recap sends a condensed version of the conversation transcript to Apple’s Private Cloud Compute system to generate the final summary.
These features are limited to the Apple Watch Series 12 and Watch Ultra 4, which must be paired with a compatible iPhone running Apple Intelligence and Siri AI (iPhone 16 and newer, excluding the iPhone 16e). Currently, the features only support English and are unavailable to users under 13. They will not initially launch in the European Union because Siri AI is not yet available there. The features are currently available in the first watchOS 27.3 developer beta, with a public beta expected before a wide release.
TYPE: news
