Apple Watch adds audio transcription and ambient conversation summaries
Apple is introducing watch features that process nearby audio and turn recent speech into text or summaries. The company says the watches will not create or store raw audio recordings. Instead, Audio Intelligence uses on-device AI to recognize sounds such as sirens, alarms, doorbells and crying babies, including when the wearer’s iPhone is elsewhere. Apple presents this function as potentially useful for people who are deaf or hard of hearing.
Two other features raise broader privacy questions. Live Rewind lets users press the Digital Crown twice to retrieve and transcribe the previous 15 seconds, with the text optionally saved in Apple’s standalone Siri app. Siri Recap can listen to conversations when enabled and produce a title, summary and key points rather than a verbatim transcript. Apple says users decide when Siri Recap is active and can switch it off through Control Center.
According to Apple, raw audio is inaccessible even to the company, speakers are not identified, and transcripts and recaps receive end-to-end encryption. The features nevertheless create questions about consent, the possible legal use of text-only records and whether people will change their behavior when recording can happen through a subtle watch gesture.
Written by Sakhanda Wire AI from reporting by TechCrunch. Facts, quotes and full context are in the original article below.