14:06Machine-translated from Azerbaijani by AIRead in Azerbaijani
Apple tests conversation summaries and transcription of the last 15 seconds on Apple Watch
Apple has released its AI-powered Audio Intelligence features for Apple Watch in beta. Siri Recap and Live Rewind have been added to the watchOS 27.3 beta released for developers. Announced in September, these features work only on Apple Watch Series 12 and Apple Watch Ultra 4 models. Public beta testing will begin later.
The system has 4 features that users can activate on demand. The sound recognition feature distinguishes sounds such as sirens, alarms, doorbells and babies crying. This may help people with hearing difficulties. The music recognition feature uses Shazam to identify the title and artist of music playing in the background. The other 2 features briefly summarize conversations and display the part the user did not hear as text. Live Rewind transcribes the last 15 seconds of a conversation. Users can view that segment by pressing the Digital Crown 2 times. They can also ask Siri questions about the text or save it to review later. An unsaved segment is automatically deleted after 30 seconds. When the feature is activated, the watch emits an alert even if it is in silent mode, and an animation and microphone icon appear on the screen. Siri Recap does not transcribe every word of a conversation; it creates only a short heading and key points. The summary may include dates, deadlines, prices and tasks, but it does not identify the speakers. Notes can be sent to the Notes and Journal apps. Unsaved summaries are automatically deleted within 7 days. At the end of the day and when the watch is turned on in the morning, the system reminds users to review these summaries. The feature can be activated manually and scheduled for specific times. According to Apple, the system does not record or permanently store raw audio. Siri Recap does not include financial and identity information, account login details or content encouraging self-harm in summaries. Conversations are not summarized while music is playing or noise cancellation mode is enabled. Users can disable each feature separately. A scheduled summary can also be stopped from Control Center. Data processing is divided among the S11 chip in Apple Watch, the connected iPhone and the Private Cloud Compute cloud system. The watch detects the conversation, and the raw audio is processed in protected memory. The phone removes extraneous words and condenses the text. The encrypted text is then sent to the cloud to generate a summary. The data in the cloud is deleted immediately after the result is returned. The operating system and apps cannot access the raw audio. Notes are transmitted between devices in encrypted form. Apple also cannot view those notes. The features require a compatible iPhone with Apple Intelligence enabled. The list includes iPhone 16 and newer models, except for iPhone 16e. Siri Recap and Live Rewind will initially work only in English. The features will not be available in the European Union and will have a daily usage limit. Apple claims that these capabilities will not affect the battery life of the watch or phone.
The system has 4 features that users can activate on demand. The sound recognition feature distinguishes sounds such as sirens, alarms, doorbells and babies crying. This may help people with hearing difficulties. The music recognition feature uses Shazam to identify the title and artist of music playing in the background. The other 2 features briefly summarize conversations and display the part the user did not hear as text. Live Rewind transcribes the last 15 seconds of a conversation. Users can view that segment by pressing the Digital Crown 2 times. They can also ask Siri questions about the text or save it to review later. An unsaved segment is automatically deleted after 30 seconds. When the feature is activated, the watch emits an alert even if it is in silent mode, and an animation and microphone icon appear on the screen. Siri Recap does not transcribe every word of a conversation; it creates only a short heading and key points. The summary may include dates, deadlines, prices and tasks, but it does not identify the speakers. Notes can be sent to the Notes and Journal apps. Unsaved summaries are automatically deleted within 7 days. At the end of the day and when the watch is turned on in the morning, the system reminds users to review these summaries. The feature can be activated manually and scheduled for specific times. According to Apple, the system does not record or permanently store raw audio. Siri Recap does not include financial and identity information, account login details or content encouraging self-harm in summaries. Conversations are not summarized while music is playing or noise cancellation mode is enabled. Users can disable each feature separately. A scheduled summary can also be stopped from Control Center. Data processing is divided among the S11 chip in Apple Watch, the connected iPhone and the Private Cloud Compute cloud system. The watch detects the conversation, and the raw audio is processed in protected memory. The phone removes extraneous words and condenses the text. The encrypted text is then sent to the cloud to generate a summary. The data in the cloud is deleted immediately after the result is returned. The operating system and apps cannot access the raw audio. Notes are transmitted between devices in encrypted form. Apple also cannot view those notes. The features require a compatible iPhone with Apple Intelligence enabled. The list includes iPhone 16 and newer models, except for iPhone 16e. Siri Recap and Live Rewind will initially work only in English. The features will not be available in the European Union and will have a daily usage limit. Apple claims that these capabilities will not affect the battery life of the watch or phone.