Cliquez ici pour lire en français
Need to recall what someone said a few seconds ago, or get a quick summary of a long conversation, straight from your wrist? That’s what Apple is aiming for with a new set of AI features for the Apple Watch. Called Audio Intelligence, the toolkit has entered developer testing through the watchOS 27.3 and iOS 27.3 updates. Apple first showed it at its hardware event last September.
Four features to keep conversations on track 🎙️
Audio Intelligence bundles four options that users can switch on as they choose. They can detect important sounds, identify songs, help you pick up the thread of a conversation and take notes automatically.
Two of them stand out. Live Rewind shows the last 15 seconds of a conversation as text, which can help when you’ve briefly lost the thread. The snippet reportedly disappears after about 30 seconds unless you save it.
Siri Recap handles longer exchanges. Instead of a word-for-word transcript, it generates a summary with a title and the main points. You can read these summaries in the Siri app and save them to Apple Notes or Journal. You can also schedule them by time or location.
Limited to certain devices 📱
For now, the beta is open to a limited number of users. It only works on the Apple Watch Series 12 and Apple Watch Ultra 4, paired with an iPhone that supports Apple Intelligence, Apple’s suite of AI tools.
The features currently support English only. They are not available in the European Union, and they are reportedly off-limits to anyone under 13. Apple says a public beta will come before a wider rollout.
Privacy is Apple’s central pitch 🔐
Always-listening devices already worry many people, from smart glasses to AI-powered recorders. These gadgets can pick up the words of people who may not know they’re being recorded.
Apple seems to have anticipated that concern by presenting Audio Intelligence as a memory aid, not a recording device. According to information shared around the beta, the watch first checks that its wearer is actually taking part in the conversation. The other people in the exchange are neither identified nor named.
On the technical side, reports from the beta say Live Rewind runs on the device itself, while Siri Recap summaries are generated through Private Cloud Compute, Apple’s secure cloud processing system. The watch also signals to people nearby when it is processing a conversation.
There are other guardrails too. According to Apple, summaries aren’t generated while the wearer is listening to music through headphones or has noise cancellation on, and Siri Recap is designed to filter out certain sensitive information.
Summaries are not a reliable transcript ⚠️
Apple itself acknowledges that its AI can get things wrong. Snippets may be incomplete, speech can be misheard and some details may be summarized incorrectly.
That matters most for anyone thinking of using Siri Recap in work meetings or to remember important instructions. Treat the summaries as memory aids, not as accurate minutes you can rely on without checking.
A new frontier for conversational AI 🚀
With Audio Intelligence, Apple brings conversation summaries to a mainstream smartwatch, which could put the technology in front of far more people than specialized AI recorders ever reached.
The open question is how users, and the people around them, will react to a watch that processes conversations, even in short bursts. The beta period should show whether Apple’s privacy assurances are enough to reassure the public before any large-scale launch.
Would you be comfortable wearing a watch that can analyze and summarize your conversations? Tell us in the comments.
📱 Get our latest updates every day on WhatsApp, directly in the “Updates” tab by subscribing to our channel here ➡️ TechGriot WhatsApp Channel Link 😉
