Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

MacRumors

macrumors bot
Original poster


The Apple Watch Series 12 and Ultra 4 have new Audio Intelligence features that include Sound Recognition, Music Recognition, Live Rewind, and Siri Recap. When these features are activated, the new Apple Watches are always listening to what's going on around the wearer.

apple-watch-series-12-siri-recap.jpg

With Siri Recap, the Apple Watch listens to your conversations throughout the day and takes notes. It uses Apple Intelligence to continually generate summaries that Apple says are meant to "jog your memory" on events so you can stay present in the moment.

Apple anticipated negative reactions to a wearable that's always listening, so it has a detailed explainer on the privacy features that have been added.
  • No speaker attribution - Siri Recap will not identify and attribute speakers. Names can be included if mentioned in the conversation, but nothing is directly attributed to a specific person.
  • Omitted information - Siri Recaps will omit potentially harmful content or sensitive data, which Apple says includes language that promotes self-harm or hateful speech, financial data, government-assigned identifiers, authentication data, and certain personal identifiers.
  • Caveats - Apple warns that Siri Recap summaries can be incomplete or inaccurate, and filtering may not remove all sensitive information.
  • No transcript - Apple says Siri Recap doesn't produce a full transcript of a conversation. It's just a summary, comparable to notes one might jot down after a conversation.
  • Condensed conversations - Encrypted, condensed text versions of a conversation are sent to Private Cloud Compute. The text is less than half the length of the original transcript, and distills the conversation to its core meaning. Conversation tone, non-essential language, filler words, and redundancies are removed.
  • No audio recording - Audio is sent to the Secure Exclave on Apple Watch and iPhone for processing and analysis. It can't be accessed by the operating system, apps, the user, or Apple, and no audio is recorded or stored.
  • Private Cloud Compute - Siri Recap uses Apple Foundation Models running on Private Cloud Compute to summarize and format the final output. Data is not accessible to Apple or anyone else.
  • Opt in, not opt out - Siri Recap can be turned on in the Apple Watch app on iPhone, and set to listen on a schedule. You can have it on only at work, or set it to never listen at night. There will be a quick access toggle in Control Center for turning it on and off.
  • Auto-delete - Summaries auto-delete after seven days if not saved to the Siri app.
  • Siri app - Summaries are saved to the Siri app and can be deleted at any time from all synced devices. Saved summaries can be exported to other apps like Notes.
  • Encryption - Summaries are end-to-end encrypted when synced to your other devices via iCloud so only your trusted devices can access them.
Apple suggests that Apple Watch wearers "be mindful" of the people around them. "While audio is not recorded and stored, and no speakers are attributed, consider those around you where conversations might be private or sensitive," reads Apple's document.

Siri Recap requires one of the new Apple Watch models and an iPhone 16 or later (except the iPhone 16e). The feature is launching later this year.

Article Link: Apple Explains What Happens to Conversations Siri Recap Hears
 
I can think of exactly one context in which this “feature” is applicable: a heated argument which has devolved to the point of, “Well, that’s what you just said!”

Of course, “applicable” in that context is the opposite of “productive.”

I’m at a loss to imagine how this could possibly have made it past the alcohol-fueled spitball session where it was thought up.

b&
 
It’s really not about whether or not we trust Apple, this is a technology that normalizes our conversations being recorded. If this is on and you sit down to a meeting you are essentially recording the conversation without consent. Apple is trying to get around that by claiming it’s a summary but that doesn’t matter. If I expect we are have a casual conversation and later on you have a point form version of what I said that is still invasive.
All the security angle they try and pitch is doing is to privacy wash their invasion of user privacy and the normalization of non-consensual recording.

Edit: Apple deserves all the backlash this will cause.
 
"Siri Recaps will omit potentially harmful content or sensitive data"

And who gets to define "potentially harmful" or "sensitive"?

This is all to normalize the behavior of 24x7 recording everything everywhere. It's truly disheartening.
My problem with it is that it just another tool to further erodes direct human to human social interaction/communication. It's ok to ask someone to repeat themselves or clarify something. If someone missed or didn't understand something the first time we're going to rely on a cold translation with no emotion or social cues? In a society with declining emotional intelligence and critical thinking skills? Neat.
 
There is no expectation of privacy in a pubic gathering. I am careful what I say in public even with the person I am speaking with. As such, I am not bothered by this feature. This kind of thing has been happening for decades, with or without an electronic device.
"It happens all the time, therefore we should embrace it happening more often, by more people, in more places, in a far easier manner."

*You* may not be bothered by this feature, but there are many *others* who are bothered by it and who do not welcome it.
 
Sure Apple's implementation is fairly reasonable - although we'll still need to write a procedure at work that it must be disabled as the military won't be interested in Apple's promises when it comes to people discussing sensitive things in front of listening devices.

The biggest worry is when everyone else inevitably copies it in around a year, but without Apple's level of privacy protections.

The other worry is when governments like the UK start wading in demanding that there ARE transcripts and that they have access to them for "child safety". Don't get me wrong, catching people harming kids would be wonderful. But AI gets things wrong.. a lot. And there are an absolute ton of privacy issues

Things are going to be full scale 1984 sooner rather than later.
 
Really creepy, unnecessary feature, possibly making the position of the guy who is in charge of this, superfluous. "What's that chime I hear?" "Oh, nothing, just an alarm for my next appointment. What were saying about your STDs again?"

Ok but if you read the article you'll see it's a non-issue. There is no audio recording, processing is all on device, it only shows text, and it only lasts for 30 seconds. The person can't save what the watch heard.
 
My problem with it is that it just another tool to further erodes direct human to human social interaction/communication. It's ok to ask someone to repeat themselves or clarify something. If someone missed or didn't understand something the first time we're going to rely on a cold translation with no emotion or social cues? In a society with declining emotional intelligence and critical thinking skills? Neat.

Right!?
And especially their example of a "waiter".
That person is literally there to tell you the specials and wait on you.

They are not going to be bothered in the slightest if you ask them to repeat something or clarify.

That's.. why they are standing there!

Can we do ANYTHING without involving our g-damn devices???
 
"It happens all the time, therefore we should embrace it happening more often, by more people, in more places, in a far easier manner."

*You* may not be bothered by this feature, but there are many *others* who are bothered by it and who do not welcome it.
We have been talking with each other in public for millions of years. All of a sudden, Apple announces a feature that mankind has been initiating naturally for thousands of years and now people on MR have a problem.

If I am in a public gathering and run my mouth without thinking, I am not going to be mad if I am later held accountable for my actions. That is how real life and personal responsibility work.

People here are upset over what some other person might do with what you (plural use) say in public. You can't control others but, you can control what you (plural use) do and say.

You (direct use) want to regulate away personal responsibility? That is the core position you are taking.
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.