Apple Watch Series 12 and Ultra 4 now support Siri Recap and Live Rewind, two new Audio Intelligence features available in beta that summarize conversations and convert audio to text.
By Adam Makins·
Apple is rolling out two fascinating new Audio Intelligence features for its latest wearables: Siri Recap and Live Rewind. If you own an Apple Watch Series 12 or Ultra 4 and don’t mind running beta software, you can test these capabilities now through the watchOS 27.3 beta. Both features lean heavily on on-device processing and Apple Intelligence to deliver real-time conversational assistance without storing your audio.
Siri Recap generates high-level summaries of conversations, complete with a title and key discussion points. It’s useful when you need a quick memory jog about what was discussed. Access these summaries directly from the Siri app whenever you need them.
Live Rewind works differently. It converts the last 15 seconds of audio into text, accessible by double-pressing the Digital Crown. This is genuinely practical for those moments when an announcement at the airport flies past you, or you zone out during a presentation. You can ask Siri about the text or save it to review later.
The appeal here is obvious. Our watches are already on our wrists, always present and ready to help. Adding conversational AI to these devices feels like a natural evolution for tech that’s already deeply integrated into our daily lives.
Here’s the elephant in the room: always-on microphones make people nervous. And rightfully so. But Apple’s approach here deserves attention.
The company claims these features don’t record or store audio. Instead, the S11 chip processes audio in isolation, then immediately deletes it. The operating system can’t access raw audio, apps can’t access it, users can’t access it, and neither can Apple. This is different from features that actually need to store data.
Apple provides additional transparency measures. When Live Rewind activates, a chime sounds and the display shows an animation with a microphone indicator, even if your watch is on silent. For Siri Recap, the summaries supposedly won’t include sensitive information like financial details.
Users retain control too. Each Audio Intelligence feature can be toggled on or off individually. This isn’t one of those situations where opting out requires digging through menus.
Naturally, there are constraints. You’ll need an Apple Watch Series 12 or Ultra 4 paired with an iPhone 16 or later that has Apple Intelligence enabled (the base iPhone 16e doesn’t cut it). Initially, Siri Recap and Live Rewind only work in English. Europeans won’t get these features at launch, likely due to regulatory considerations. There are also daily usage limits, though Apple hasn’t detailed what those look like yet.
These features also don’t identify or attribute speakers in conversation. So if three people are talking, your watch won’t tell you who said what. It’s a reasonable privacy trade-off, though it does limit the usefulness in group settings.
What’s interesting is where this fits in Apple’s larger strategy. The company has been gradually moving AI processing onto devices rather than relying on cloud computation. This reduces data transmission and addresses privacy concerns head-on. Siri Recap and Live Rewind exemplify this philosophy.
These features also highlight why Apple’s wearables command premium prices. The S11 chip isn’t just capable; it’s designed specifically to handle on-device processing tasks that cheaper competitors can’t match. That specialized hardware matters when privacy is the selling point.
Of course, skepticism is warranted. Trust in tech companies around data privacy has eroded significantly. But Apple’s architecture here, where audio is processed and deleted in isolation without any copy reaching the operating system, is architecturally sound. Whether it works as claimed in practice remains to be seen once the features roll out publicly.
The fact that Live Rewind announces itself with a chime and visual indicator suggests Apple understands the surveillance concerns and is building in accountability measures.
These features won’t revolutionize how we use smartwatches. But they do represent a thoughtful approach to adding AI capabilities while respecting user privacy, which is increasingly rare in this space.
Source: Engadget
If Apple can genuinely deliver conversational AI on wearables without centralizing our audio data, does that become the new baseline privacy expectation for the entire industry?
