What Apple has put into beta
Apple has started testing Audio Intelligence, a set of opt-in AI features for the Apple Watch, with developers. The Verge describes the release as part of the watchOS 27.3 developer beta, while TechCrunch ties it to the iOS 27.3 update. The company first announced the technology at its Surprise and Shine hardware event in September. According to TechCrunch, the news first surfaced through a press release from Apple's communications team in India early on Tuesday, and the rollout reached the US and other markets later that day.
TechCrunch reports four features in total, aimed at Apple Watch Series 12 wearers: detection of important sounds, song identification, catching up on conversations and note-taking. Sound recognition is meant to help deaf and hard-of-hearing users notice sirens, alarms or doorbells, and song recognition uses Apple's Shazam app to name background music and its artist. The two features that listen to people talk are Live Rewind and Siri Recaps.
How Live Rewind and Siri Recaps work
Live Rewind is meant for moments when a user missed something just said. A double-press of the Digital Crown displays the last 15 seconds of audio as a text snippet on the watch. TechCrunch says a chime and an on-screen animation play at that moment, so people nearby have a cue that something was captured. The user can ask Siri about the content, or save the snippet to the standalone Siri app. A snippet that is not saved is deleted automatically after 30 seconds, according to TechCrunch.
Siri Recaps works over a longer span. With it switched on, Apple Intelligence generates a title and key points in the Siri app after a conversation, as The Verge reports. TechCrunch adds that Apple says the feature does not identify speakers, keep raw audio or give verbatim transcripts. Bullet points may cover dates, deadlines or pricing, and Apple suggests uses such as work meetings, parent-teacher conferences and personal training sessions.
Users are prompted to review recaps at the end of the day and again when they first unlock the watch in the morning, TechCrunch reports. Saved items can go to the Notes app or the Journal app, and anything left alone is deleted within seven days. The feature can run manually or on a schedule, such as work hours only, and can be turned off from Control Center. It does not record while the user listens to music on headphones or uses noise cancellation. Apple says it is designed to leave out sensitive material, including government identifiers and financial or authentication data.
Privacy design and hardware requirements
Apple has published an Audio Intelligence Privacy Overview, The Verge notes, addressing concerns about a watch that listens continuously. Apple says raw audio stays in dedicated hardware, is never stored as a file, and is out of reach of the operating system, apps and Apple itself. The Verge also relays Apple's statement that saved recaps and snippets sync end-to-end encrypted across devices for users with iCloud two-factor authentication and a device passcode.
TechCrunch gives a more detailed account of the pipeline. Small AI models on the S11 chip detect whether the wearer is part of a conversation, and the audio goes to the paired iPhone through a secure pairing mechanism. The iPhone converts it to condensed text with filler words removed. That text is encrypted and sent to Apple's Private Cloud Compute to build the recap, which Apple says is deleted from the service right after it is returned. TechCrunch lists the Apple Watch Series 12 and Apple Watch Ultra 4 as the supported models, and says Apple expects no effect on battery life for the watch or phone.
TechCrunch frames the conversation-listening features as potentially controversial given the backlash to AI recording devices such as Meta's AI glasses, and says Apple is trying to make them useful without ignoring the privacy of bystanders.
Health app and what comes next
The Verge reports that Apple is also testing a redesigned Health app in the iOS 27.3 developer beta. It adds an Insights tab with summaries of heart, sleep, fitness and other data, a For You section with personalised recommendations, and educational videos and articles. Apple also announced expanded language support, including global English, Simplified and Traditional Chinese, French, Japanese, Korean, Norwegian, Portuguese, Spanish, Turkish and Vietnamese.
On timing, TechCrunch says the Audio Intelligence features will move to the public beta for wider testing before reaching all users. Neither source gives a date for general availability.