Apple Watch’s AI features for rewinding and recapping conversations begin beta testing

Apple's AI-powered Audio Intelligence features for the Apple Watch have entered beta testing, initially for developers in the iOS 27.3 update, after being announced at Apple's September hardware event. The suite includes four opt-in features for Apple Watch Series 12 wearers: sound recognition for deaf and hard-of-hearing users, Shazam-based song recognition, Live Rewind, and Siri Recaps. Live Rewind lets users double-press the digital crown to see a text snippet of the last 15 seconds of speech, accompanied by a chime and animation to signal recording; snippets are deleted after 30 seconds unless saved to the Siri app, and users can ask Siri follow-up questions about the captured text. Siri Recaps summarizes conversations into a short sentence and bullet points of key facts without identifying speakers, saving raw audio, or producing word-for-word transcriptions, and is designed to omit sensitive information such as government identifiers, financial and authentication data, and language promoting self-harm. Recaps prompt users to review and choose what to save to Notes or Journal, and self-delete within seven days if untouched. The features rely on the S11 chip, built-in microphone, a secure enclave audio buffer, on-device iPhone processing, and Apple's Private Cloud Compute, with audio inaccessible to Apple and recaps encrypted across devices. Apple says the features should not impact battery life, with a public beta to follow.

(Source：TechCrunch)