Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

More Android phones than ever are covertly listening for inaudible sounds in ads

arstechnica.com

Post reply on HN
Print view View original

101–107 of 107 posts

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#101

Post by pjmlp »

on_and_off wrote: post #100
pjmlp wrote: post #92

Earlier quoted context omitted.

Not to Google, as they don't care. Now with Android N reaching 7%,one year later, who do they think will even bother with Android O?

I work on a large Android app and we do bother with Android O. As soon as the sdk will be final we will think about compiling against it and targeting it. We are not in a huge hurry, it can skip a release or 2 (we release every 3 weeks). There are no big breaking changes though (unlike let's say Marshmallow) so as long as the support lib is stable (it needs to match with compile version) we are going to support it AS…

How many customers do you feel will have devices with Android O, given Android N is around 7% after one year?

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#102

Post by beejamin »

Is there any legitimate need for a phone mic to pick up sounds way outside the range of human speech and hearing? If this sort of thing should not be allowed, why not just add a software low-pass filter in the front of the audio input chain? What would the side effects of this be?

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#103

Post by on_and_off »

pjmlp wrote: post #101
on_and_off wrote: post #100

Earlier quoted context omitted.

I work on a large Android app and we do bother with Android O. As soon as the sdk will be final we will think about compiling against it and targeting it. We are not in a huge hurry, it can skip a release or 2 (we release every 3 weeks). There are no big breaking changes though (unlike let's say Marshmallow) so as long as the support lib is stable (it needs to match with compile version) we are going to support it AS…

How many customers do you feel will have devices with Android O, given Android N is around 7% after one year?

It goes further than that.

First, that 7% figure is the whole Android (with play services) user base, numbers are pretty different on the dashboard of our general public app.

Secondly, we will start by compiling for O : the supports libraries are only tested for their corresponding compile version. That way we will be able to use new support lib features on all Android versions.

Then, we will target O (probably in the same release of the app). It should be pretty trivial : this is not marshmallow with the new permission system, system level changes in O are manageable.

For clarification, the build script of an android app separates compile version (= binary compatibility) and target version (= you handle the new behaviors of the system like granular permission in M)

By targeting O ASAP we :

- are potentially able to ship some features based on this release, like shortcuts for N. Sure, we are not going to spend a lot of dev time on an O-only feature right now but there are some easy wins. And of course the install base does not stay small for very long.

-are able to detect potential problems before OEMs start launching flagships with that Android version and we get millions of crashes / day.

There is really no good reason to stay behind.

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#104

Post by Markoff »

voidz wrote: post #62
Markoff wrote: post #47

Earlier quoted context omitted.

that's quite amusing suggestion considering Greenify is closed sourced Chinese root app which optionally (only?) report data back to mainland China and dev worked for some companies producing shady apps, before according his LinkedIn profile

Could you back this up / substantiate your claim, please?

there is option to send some statistics data back to developer, it's opt-in

I am not saying he is actualy sending some other data without permission of user, but being closed sourced Chinese ROOT app is more than enough to stay away from it in days of Nougat/Marshmallow making this app pretty much useless anyway.

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#105

Post by on_and_off »

pjmlp wrote: post #95
on_and_off wrote: post #71

Earlier quoted context omitted.

The landscape of people who actually download apps is very different though. And he has a point : Google could easily ask that all new apps target the last version of the OS.

Just as they could "nicely" require OEMs to get their game straight.

Not sure how it could work out tbh.

Ok I am not a lawyer or a part of the negotiations between Google and OEMs but I don't think Google can force OEMs to keep phones up to date for a given amount of time.

First, it is not just up to the OEMs, they need new drivers for new platform versions.

Secondly, what leverage does Google have here ? Preventing OEMs from releasing phones with Play Services looks like a really sharp double edged sword. Even triple edged actually since consumers would also suffer from this kind of deal.

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#106

Post by blurbleblurble »

craftkiller wrote: post #99
IanCal wrote: post #90

Earlier quoted context omitted.

These are like modern day ghost stories. Do you really think they're running constant speech recognition on all phones, apparently unnoticed so far, just to recommend YouTube videos?

I agree that we should remain skeptical, however, one important distinction between this and ghost stories is this is already fully technically possible. All that's needed is the will, the raw processing power, and the battery life to support it. No one could ever create a proof of concept for ghosts.

This comment is an argument that it's possible, not a claim that it's happening.

It'd be much, much easier for apps like FB messenger to transmit sparse samples of low bitrate audio to a server for processing there. 1 hr/day @ 8kbps ~ 110 MB/month. As a matter of fact, with "deep learning" and whatall, it doesn't seem like the audio would even need to be intelligible to human ears for it to still be useful in some marketing algorithms or to generate guesses of interests correlating to the audio. Those interests might be presented as an ad, search suggestion, etc. Why would google use something like this for search recommendations? Because it's training data, that's why!

What I'm arguing is that even though conventional "speech recognition" is processor intensive and unlikely happening on a mobile client side, it's just not that far-fetched that small snips of low bitrate conversation data might be used to tune marketing prediction models on the server-side.

In training general purpose models, huge amounts of data could be generated by only a little bit of data/user. Where would this sort of thing cross a privacy line? I'm not sure. Personally, it kinda creeps me out regardless.

Seriously, listen to these 8kbps opus samples: http://www.opus-codec.org/examples/ . If they're intelligible to human ears, I can't imagine what some clever learning algorithms could do with that and much much less.

Here's a random anecdote about how much data FB messenger can use: https://www.reddit.com/r/androidapps/comments/49yhlh/faceboo...

Re: More Android phones than ever are covertly listening for inaudible sounds in ads

#107

Post by IanCal »

craftkiller wrote: post #99
IanCal wrote: post #90

Earlier quoted context omitted.

These are like modern day ghost stories. Do you really think they're running constant speech recognition on all phones, apparently unnoticed so far, just to recommend YouTube videos?

I agree that we should remain skeptical, however, one important distinction between this and ghost stories is this is already fully technically possible. All that's needed is the will, the raw processing power, and the battery life to support it. No one could ever create a proof of concept for ghosts.

Replace ghost stories with the creepy house at the end of that old road having an old woman in who eats children if they try and get their ball back, and turns them into puppets making the next victim watch the show, then. And it's totally happening because I heard a crunch one time that must have been bones.

The fact that it's possible to eat a child or make a puppet does little to make this kind of thing much more realistic.

The much simpler answer is that you are more predictable than you think, and likely more influenced by things Google can watch (clicks, ads) than you think. You also don't spot when it doesn't recommend something so accurate so it could even just be chance (or at least doesn't mean they're good at it)

Post reply on HN