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

Anonymized Siri Voice Clips Stored by Apple for Up to Two Years

I couldn't care less. If I had something truly secret I wouldn't use Siri, and probably wouldn't even use Calendars since that all gets backed up to their cloud anyway. I'd probably use some vaulted software on my phone to keep that sort of stuff.

Paranoia runs deep. On any given day, I probably show up on hundreds of surveillance cameras, dozens of cell towers, several GPS satellites, dozens of websites tracking what I do, etc. etc.

Why would I care if Apple stores my siri calendar and reminder requests (as well as the occasional goofy/profanity laced/sexually charged requests that we all have tried) especially if they have specifically disassociated it with my identity?
 
scottsjack said:
There will be two kinds of responses on this thread.

The second group will be those who avoid Siri and think outside of the walled garden. To them Apple's '1984' ad has now come true in the reverse. They either do not use social media or use it with caution. They may also avoid storing or transferring data on the cloud.

By page four the two sides will be arguing.
Click to expand...

Hehe, I'm firmly in the second camp, but it won't stop me using Siri if I find a great need for it. However I think it's mostly just a gimmick and I've only used it a couple of times.

But I'll up you and say they'll be arguing by the end of this page or page 3! :D
 
maflynn said:
My only question is why especially since its anonymized - what benefit is there to Apple?
Click to expand...

That is actually very simple. Let's say a developer at Apple working on improving Siri makes some change to make it understand things better. He or she can then run the old Siri code and the new Siri code against ten million stored voice clips and check which things are recognised differently.

Let's say Siri has problems keeping "home" and "hole" apart. A developer changes the code. Out of ten million stored voice clips 800 are now interpreted differently. So the developers then listens to these 800 clips. If "Honey, I'll be hole late" changed to "Honey, I'll be home late" that's good. If "I dug myself a hole" changes to "I dug myself a home", that's bad. If "where can I buy a comb" changes to "where can I buy a coal" due to the change, that's bad.

So this lets Apple check any potential improvements to Siri against a huge database to see if an improvement really works and doesn't have negative side effects.
 
nwcs said:
I don't see much harm if they are in fact doing what they said. It is also extremely likely Samsung and Google are doing similar with their respective services. The value in having these files, even disassociated, is that you start analyzing trends in usage as well as commands/requests made. It can guide a company on where to improve their product, add new features, or highlight uses people don't seem to be using.

I'm definitely someone who values privacy, and I also have worked in software for 20 years and know exactly how customer data is usually handled on the backend. I really don't have a problem with this. And, honestly, I have no problem with Google's data collection per se. What I dislike is advertising which is their business.
Click to expand...

I think I'm going to start making requests to Siri that I know she can't do, but she should be able to. Maybe some unpaid intern within Apple will hear it and decide it should be added :D
 
This is kinda scary I guess but all companies do this. I'm sure even iMessage is store indefinitely. Google, Amazon are no different. OH and definitely FaceBook!
 
I imagine the ones they should retain for further refinement are the ones that return the response, "I'm sorry. I do not understand."
 
scottsjack said:
There will be two kinds of responses on this thread.

The first will be "big deal". Those are the people who love Apple's walled garden, who don't mind the restrictions and the DRM. They may feel that if you have done nothing wrong you have nothing to fear. They probably use a lot of social media without much thought to privacy.

The second group will be those who avoid Siri and think outside of the walled garden. To them Apple's '1984' ad has now come true in the reverse. They either do not use social media or use it with caution. They may also avoid storing or transferring data on the cloud.

By page four the two sides will be arguing.
Click to expand...

As a mass murderer who told Siri where all the bodies are. (had to confess things to somebody, since I feel lonely and misunderstood) I can vouch that Siri is keeping our conversations private, because I have not been visited by the police or FBI (yet)

May this post will do it! ;)
 
Last edited:
Does anyone know how long Google stores their voice data?

(That includes voicemail transcripts from Google Voice, voice input from Google Now, and data from Google Actions)
 
That's just creepy. These big megacorporations have to start respecting their customers.

No SIRI for me, no friggin way.
 
paulrbeers said:
Why would I care if Apple stores my siri calendar and reminder requests (as well as the occasional goofy/profanity laced/sexually charged requests that we all have tried) especially if they have specifically disassociated it with my identity?
Click to expand...

With available technology of voice biometrics which allows speaker recognition, the clip is about as anonymous as if you had your fingerprint stored with it.
 
People want improvements to Siri. This is how you get it.
This data can be used to:
Learn how people are phrasing questions
Learn what are people asking that Siri can't answer
Improve voice recognition for sample that returned "low confidence" results.
Make Siri's own voice to sound more natural
Improve how they compress the packets sent to apple w/o compromising results
Improve mic optimization a
Improve background noise isolation.
And the list goes on…

People complain that Apple isn't doing enough R&D, but this IS how you research voice interfaces.
 
jrswizzle said:
My question would be - so what?
Click to expand...

So what depends on what is done with the data. My point was that the usual consensus is Google does this in ways that are often malicious. But we're probably going to give Apple the benefit of the doubt and say "so what" even though we don't know where it's headed.
I'll still use Google and I'll still use Siri, but I'll also be a little more interested in how Apple is using Siri data (beyond just "improving service").
I mean, how is this different than reading email? I compose texts and emails in Siri all the time. It's not just "tell me a joke" crap.

----------
err404 said:
People want improvements to Siri. This is how you get it.
This data can be used to:
Learn how people are phrasing questions
Learn what are people asking that Siri can't answer
Improve voice recognition for sample that returned "low confidence" results.
Make Siri's own voice to sound more natural
Improve how they compress the packets sent to apple w/o compromising results
Improve mic optimization a
Improve background noise isolation.
And the list goes on…

People complain that Apple isn't doing enough R&D, but this IS how you research voice interfaces.
Click to expand...

Right, but it can also be used to:
Find out what you buy
Read your email
Read your texts
Learn about your family

I'm not saying that's what happening, or that it isn't happening in other ways already, just that not everything about this procedure is benign or beneficial.
 
LOL... how different would this thread be if the title was:

"Anonymized Google Now voice clips stored by google for two years"

You guys would be flipping out!

LOL... too funny...
 
maflynn said:
My only question is why especially since its anonymized - what benefit is there to Apple?
Click to expand...

They can keep the clip, keep the interpretation that Siri gave it and see if they matched. If not, teach it.

What I want to know is what the ACLU is doing about all the footage of me at banks, malls, CCTV on the streets. Many of those can be identified based on transaction time stamps. How are they protecting my privacy with those.

Or do they only care about big name tech companies
 
You know, I'm someone who really values my privacy. I like privacy for privacy's sake, and I can't tell you how many posts, articles and letters I've written on the subject. You know what, this doesn't bother me at all, and not just because it's Apple.

Rather than the fact that it's a company I like that's doing it this time, it's actually more because of the fact that the information is of absolutely no value whatsoever. Maybe you're different, but as much as I love privacy, I don't care of someone listens to an anonymized clip of me saying 'shuffle my music' all day long. I can't think of anything I've said to Siri that would be even remotely compromising, let alone identifiable on its own.

The last thing I want to do is take the 'I've got nothing to hide approach,' I can't stand that. But the fact is, this is no different than a search engine storing your search history. It's actually quite similar.

And I know I said this wasn't about Apple, but really, I'd rather have Apple storing my data than Google. At least it's anonymized, and they have less incentive to tie it to me and figure out ways to monetize it.

If I do catch Apple doing something I think is creepy, I'll be the first to quit using the service and speak out against it.

– Dictated with Siri
 
Siri relies on previous data to make judgement calls on the tasks that it does, ie it learns. It's obvious to any Siri user that this is the case, as Siri often recalls previously discussed tidbits of information during use. If you're uncomfortable with Siri having this info, then you can either not use it or turn it off, so I don't know why anyone would be "concerned" about this non-issue.
 
maflynn said:
My only question is why especially since its anonymized - what benefit is there to Apple?
Click to expand...

Apple uses the recordings as sample data to help improve Siri's natural language recognition. Converting the recording to text is only half the job. (Which is handeld by Nuance's technology) Siri now needs to figure out what this text means. Having a large sample of different types of slang and and regional dialects helps greatly.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back