Apple Normalizes Always-Listening Hardware With New Watch AI

Apple is attempting to shift consumer expectations regarding privacy with the introduction of "audio intelligence" tools on the new Apple Watch Series 12 and Ultra 4, according to reporting from TechCrunch and Wired.

Introduced at the Surprise and Shine event, the new features include "Live Rewind," which allows users to double-press the Digital Crown to transcribe the previous 15 seconds of environmental audio, and "Siri Recap," which uses ambient listening to generate high-level summaries and key points of conversations. While Apple suggests these tools help users remember book recommendations or meeting takeaways, TechCrunch notes that the ability to subtly record others raises significant ethical and legal questions regarding consent.

To mitigate privacy concerns, Apple told Wired that the features do not store raw audio recordings. The company utilizes a memory-protected space called the "Secure Exclave" on its new S11 chips to process audio data in an isolated buffer. For Siri Recap, audio is encrypted and transmitted to a paired iPhone, where local models transcribe the speech before the raw audio is deleted. The resulting condensed transcript is then processed via Apple's Private Cloud Compute to create a summary, which Apple says is designed to strip out personal identifiers like ID numbers or financial data.

Other tools include Sound Recognition, which alerts users to sounds like sirens or crying babies, and an updated Shazam feature. While Apple emphasizes that Siri Recap is opt-in and can be scheduled or toggled via the Control Center, TechCrunch reports that the normalization of such listening technology may clash with broader consumer backlash against surveillance tools like AI apps and Flock cameras.