Apple is testing new artificial intelligence (AI) features for the Apple Watch Series 12 and Apple Watch Ultra 4 that could change how people remember conversations and manage everyday information. The company’s Audio Intelligence technology includes Live Rewind, which displays recently spoken words as text, and Siri Recap, which creates short AI-generated conversation summaries. Apple first introduced these capabilities in Apple’s September 2026 announcement, and developer beta testing has now begun.
Developer beta testing of Live Rewind and Siri Recap began in early November 2026, with the features arriving through the iOS 27.3 and watchOS 27.3 beta updates. Apple originally announced these conversation-related AI features in September 2026. The new beta allows eligible developers to test the technology before a wider release. However, Live Rewind and Siri Recap are not yet generally available to all Apple Watch owners, and Apple has not confirmed an exact public launch date.
Live Rewind allows users to check the last 15 seconds of spoken conversation when they miss or misunderstand something. By double-pressing the Apple Watch’s Digital Crown, users can see a short text snippet showing what was recently said. They can ask Siri about the text or save it in the Siri app for later reference. Apple says Live Rewind does not continuously create transcripts in the background. If users do not save or ask Siri about the snippet, the displayed text disappears permanently approximately 30 seconds after the Apple Watch screen dims.
Siri Recap is another major Audio Intelligence feature designed to help users remember important conversations. When enabled, it uses AI to create a high-level summary rather than a complete word-for-word transcript. Each recap can include a title, a short summary, and important points that users can review in the Siri app on their Apple Watch or paired iPhone. For example, someone could use it to remember tasks discussed during a work meeting, details from an appointment, or plans made with friends. Apple says Siri Recap does not identify individual speakers or label who said particular words, helping reduce the amount of personal information included in summaries.
Audio Intelligence also includes Sound Recognition and automatic Music Recognition with Shazam. Sound Recognition can detect important noises such as alarms, sirens, doorbells, and crying babies, then send notifications to the Apple Watch. The feature is designed to help people who are deaf or hard of hearing and can work even when the paired iPhone is not nearby. Meanwhile, Shazam can automatically identify music playing in the background and display the song title and artist in the Apple Watch’s Smart Stack without requiring users to open the app manually. These sound and music features are separate from the conversation-related Live Rewind and Siri Recap beta capabilities.
Privacy is one of the biggest considerations for Apple’s new conversation-related AI features because the Apple Watch can process spoken information nearby. According to Apple, Audio Intelligence does not create or store audio recordings. Instead, the S11 chip uses a protected hardware system called Secure Exclave to process audio separately from ordinary apps and operating system functions. Apple says the raw audio is deleted after processing and cannot be accessed by Apple or other apps. When someone activates Live Rewind, the watch also plays an audible chime and displays an animation with a microphone indicator. These signals are intended to alert nearby people that the feature has been activated, even when the watch is set to silent mode.
Apple says users can control when its conversation-related AI features are active. Siri Recap can be turned on or off manually through Control Center, while schedules can limit its operation to particular times or locations, such as working hours. Users can review, edit, save, export, or delete summaries in the Siri app, and unsaved Siri Recap summaries are automatically deleted after seven days. Live Rewind requires a deliberate double-press of the Digital Crown and does not continue transcribing after the request. For detailed settings, deletion rules, and privacy options, users can consult Apple’s Audio Intelligence support guide.
Apple’s Audio Intelligence technology combines processing on the Apple Watch and paired iPhone with Private Cloud Compute for certain AI tasks. For Siri Recap, spoken information is protected as it moves from the watch to the paired iPhone, where it is transcribed and condensed in a secure hardware environment. An encrypted, shortened version of the conversation text is then sent to Private Cloud Compute to generate a summary, without sending the original raw audio recording to the cloud. Apple says saved Siri Recap summaries and Live Rewind text snippets are protected with end-to-end encryption when synced through iCloud on properly configured devices. More technical information about this cloud security approach is available in Apple’s Private Cloud Compute security documentation. However, AI-generated summaries can still misunderstand conversations or omit important details, so users should verify critical information with the people involved.
Live Rewind and Siri Recap require an Apple Watch Series 12 or Apple Watch Ultra 4 paired with a compatible Apple Intelligence-enabled iPhone. Supported phones include the iPhone 16, iPhone 16 Plus, iPhone 16 Pro, iPhone 16 Pro Max, and eligible newer models, but not the iPhone 16e. Users must also have Apple Intelligence and Siri AI enabled. The conversation features initially support English, with more languages planned for the future, and will not be available in the European Union at launch. Apple also states that Live Rewind and Siri Recap are unavailable to children under 13, while users under 18 with Child Accounts require parental permission. Certain AI processing features may also have daily usage limits. Sound Recognition and automatic Shazam Music Recognition have less demanding iPhone requirements, supporting compatible Apple Watch models paired with an iPhone 11 or later, or a second-generation iPhone SE or later, running iOS 27.
These new capabilities show how Apple is expanding the role of its smartwatch beyond fitness tracking, health monitoring, and notifications. The developments also connect with Apple’s broader AI wearables plans, which explore how personal devices could use more advanced Siri interactions and contextual assistance. Features such as Live Rewind and Siri Recap may help professionals, students, and everyday users manage information more easily, although accuracy and privacy will remain important considerations.
Apple has not confirmed an exact public release date for Live Rewind and Siri Recap. Their beta testing comes as the company continues expanding artificial intelligence across its devices and services. TechBonafide previously covered Apple’s partnership with Google Gemini for AI-powered Siri, which provides background on the company’s broader AI strategy. However, Apple describes Siri Recap as using Apple Foundation Models and has not specifically confirmed that Gemini powers this Apple Watch feature. For now, the success of Audio Intelligence will depend on how accurately it summarizes conversations, how useful its controls are, and whether users feel comfortable enabling conversation-related AI features in daily life.
