Apple Watch's new AI listening features spark privacy debate
Apple has added AI-powered audio processing and transcription features to its new watches, letting them detect sounds and summarize conversations. The move offers useful accessibility benefits but also raises questions about consent and constant listening.
At its recent hardware event, alongside a foldable iPhone, Apple unveiled new Apple Watch features built on on-device AI that can process nearby audio and turn spoken words into text. The move marks a shift for a company long known for emphasizing privacy, as it edges toward technology that is effectively always listening to its surroundings.
Three new capabilities
The update introduces three features. "Audio Intelligence" uses on-device AI to detect sounds such as sirens, alarms, doorbells, and crying babies, alerting the wearer — a feature aimed at helping deaf or hard-of-hearing users, which also works without a paired iPhone nearby.
More contentious are "Live Rewind" and "Siri Recap." Live Rewind lets a user double-press the Digital Crown to rewind the last 15 seconds of audio and transcribe it, saving the text to Apple's new standalone Siri app. Siri Recap, meanwhile, listens ambiently during a conversation and generates a title, summary, and key points afterward rather than a full transcript.
Questions over consent and trust
Unlike some competing AI wearables, these features are not on by default — users can choose when to enable them, such as only during work hours, and can switch them off anytime from Control Center. Apple says it does not create or store raw audio recordings, that audio isn't accessible even to Apple itself, that speakers aren't identified, and that transcripts and summaries are protected with end-to-end encryption.
Still, observers point out the implications of the design. Unlike taking a photo or video, which requires a visible gesture, capturing audio via a quick button press is far more subtle, even though it triggers an audible chime and an on-screen animation. The functionality raises questions about consent from the people being recorded, how such transcripts might be treated as evidence in legal proceedings, and how the awareness of potential recording could shape everyday behavior and conversation.
