ARTIFICIAL INTELLIGENCE · FOCUS
Siri Recaps tests the limits of automated audio memory
Apple is testing audio features that can transcribe or summarize conversations around the watch. A review published by TechCrunch highlights Siri Recaps' usefulness, but also its tendency to retain trivial details and mix different subjects.
Legal disclosure: this content was produced by an artificial intelligence system. Editorial transparency
On Tuesday, Apple introduced Audio Intelligence features for Apple Watch Series 12 users in the first developer betas of iOS 27.3 and watchOS 27.3, TechCrunch reports. They include detection of important sounds, music identification and tools for remembering conversations.
Live Rewind lets users review what was said during the past 15 seconds. Siri Recaps produces notes summarizing information deemed important in the day's audio. The review published by TechCrunch describes useful applications during arrangements involving a car repair and a rental: the system retained practical information communicated verbally.
The same review, however, identifies limitations in selection and organization. Vehicle information was grouped with family conversations, while unremarkable television viewing plans were also remembered. This individual account is not a general measure of reliability, but it illustrates the difficulty of automatically deciding what deserves a note.
According to TechCrunch, Apple plans to process information on the device when possible, without identifying speakers or retaining raw audio files, while masking certain private data. These technical protections leave another question about usage open: when does the user want their conversations to become memories they can consult?