Please create an account to participate in the Slashdot moderation system

 



Forgot your password?
typodupeerror
×
Privacy Android Communications IOS Security

Why Phones That Secretly Listen To Us Are a Myth (bbc.com) 219

A mobile security company has carried out a research investigation to address the popular conspiracy theory that tech giants are listening to conversations. From a report: The internet is awash with posts and videos on social media where people claim to have proof that the likes of Facebook and Google are spying on users in order to serve hyper-targeted adverts. Videos have gone viral in recent months showing people talking about products and then ads for those exact items appear online. Now, cyber security-specialists at Wandera have emulated the online experiments and found no evidence that phones or apps were secretly listening. Researchers put two phones -- one Samsung Android phone and one Apple iPhone -- into a "audio room". For 30 minutes they played the sound of cat and dog food adverts on loop. They also put two identical phones in a silent room.

The security specialists kept apps open for Facebook, Instagram, Chrome, SnapChat, YouTube, and Amazon with full permissions granted to each platform. They then looked for ads related to pet food on each platform and webpage they subsequently visited. They also analyzed the battery usage and data consumption on the phones during the test phase. They repeated the experiment at the same time for three days, and noted no relevant pet food adverts on the "audio room" phones and no significant spike in data or battery usage.

This discussion has been archived. No new comments can be posted.

Why Phones That Secretly Listen To Us Are a Myth

Comments Filter:
  • bullsh!t (Score:5, Informative)

    by sdinfoserv ( 1793266 ) on Friday September 06, 2019 @11:16AM (#59165688)
    When you say "Ok Google/Siri" and the phone "wakes up" and responds.. it's frigging listening - always- period.
    • Re: bullsh!t (Score:4, Insightful)

      by spcebar ( 2786203 ) on Friday September 06, 2019 @11:24AM (#59165730)
      There's a bit difference between monitoring speech patterns to detect a wake word and scraping your conversations for advertising purposes.
      • Re: bullsh!t (Score:5, Interesting)

        by sexconker ( 1179573 ) on Friday September 06, 2019 @11:53AM (#59165840)

        Buy an Alexa.
        Plug it in and set it up.
        Say "Blarble goo ha wiz wozzle".
        Wait 5 minutes.
        Log into your Amazon account on a PC.
        Dig through the devices pages to find your Alexa.
        Dig through the settings/troubleshooting pages.
        Find the page that lists what Alexa has recorded.
        See that Amazon helpfully lists an instance of speech detection that was not meant for Alexa (because it didn't hear "Alexa").
        See that you can download and listen to that instance.
        See that "Blarble goo ha wiz wozzle" was recorded and sent to Amazon and is available for people on the internet to download and listen to.

        • by ArmoredDragon ( 3450605 ) on Friday September 06, 2019 @12:55PM (#59166088)

          First of all, that's an "Alexa" that you bought, not a smartphone. Second of all, it's pretty easy to detect speech without actually recording anything beyond making a log entry stating that somebody said something.

          And you do have to ask, at some point, where you're going to draw the line on what's listening to you and what isn't. Though, if it isn't transmitted to any other entity and never leaves your vicinity, then what difference does it make?

          Of course, virtually every computing device we use has the capability of recording audio in some capacity, and it doesn't matter whether you have a microphone connected to it. If it has any kind of precision instrument, like a hard disk or an accelerometer, it can, in practice, be used to record audio. In some cases, that also includes capacitors, which are in basically everything.

          You know what that means, right? We, the Illuminati Order, can track every move you make, and every sound you make, and therefore predict very accurately what you'll do in the near future. We know you better than you do. And there's nothing you can do to stop us.

        • Re: bullsh!t (Score:4, Insightful)

          by thegarbz ( 1787294 ) on Friday September 06, 2019 @02:37PM (#59166552)

          I just said "Blarble goo ha wiz wozzle", and 5 minutes later logged into Amazon. The only thing that is helpfully listed is the recording purposefully made immediately following the most recent time I said "Alexa".

        • Why would you even post this? It's easily checked. Worse yet, what fools modded him up?

          For those that don't actually want to bother, the answer is no. The entire post is garbage. You won't see anything you said without a wake word.

      • If it's not scraping your conversations for ads, it's at least priming the voice recognition, which shares data with the search bar. Everytime I'm talking to my mother about something and I decide to search related to it, the autocomplete knows exactly what I want, even if I've never had an interest in that topic before.

    • Re:bullsh!t (Score:5, Insightful)

      by jellomizer ( 103300 ) on Friday September 06, 2019 @11:24AM (#59165732)

      It is hearing, not listening.

      Normally the Ok Google/Hey Siri call out isn't sent to the cloud, but in the software on the OS. Voice Recognition for a few simple words, can be done in hardware. What you say after that is sent to the cloud and back with a data response. As it needs more computing power to interpret what you actually said.

      Now listening is comprehending at some level. When you talking there is no comprehending other than you phone is going does this waveform match Ok Google or Hey Siri wave form with appropriate deviation.

      You can put your phone in a Faraday cage and try it out, see what works and what doesn't when there is no radio connection.

      • Re: (Score:2, Interesting)

        You SERIOUSLY think that the phrases "hey siri" or "ok google" are the only phrases that "activate" the device to further "listen" to you? It's certainly demonstrable that other phrases cause adverts of a specific type. We all know that if there's no direct law against it, then some corporation is going to exploit everything possible. Why not this?

        • Re: bullsh!t (Score:4, Insightful)

          by Iconoclysm ( 3885655 ) on Friday September 06, 2019 @12:15PM (#59165926)
          If itâ(TM)s demonstrable why hasnâ(TM)t it been demonstrated?!
        • Re: (Score:3, Informative)

          by swillden ( 191260 )

          It's certainly demonstrable that other phrases cause adverts of a specific type.

          Is it? This is exactly what this experiment tried to demonstrate, and failed. Do you have a link to some other demonstration that succeeded?

          • Re:bullsh!t (Score:5, Informative)

            by BringsApples ( 3418089 ) on Friday September 06, 2019 @03:17PM (#59166766)

            This BS test that they did, they didn't have casual conversations, they put the stupid phones in a room where adverts for cat and dog food was looped. This test was presented to you by the same people that are listening to you. It was a BS test, with BS results.

      • I'm not an expert, but I did read something about this written by a guy who claimed to be an expert. According to them one if not the reason the speech is sent up to the cloud is to protect the speech recognition software from being torn apart and analysed. His assertion was that training the various parts of the recognition ware was very expertise and resource heavy but that RUNNING it afterwards was not. Not sure how true that is but it sounds plausible for a huge service that has to scale out.
        • I'd be willing to bet it could be reasonably run locally. I think image recognition would be more difficult than speech recognition, and google's already made some pretty good progress in getting that to work offline.

          I've got a pixel 3 will the Playground camera mode (ie: altered reality). The following worked shortly after it came out (but Playground doesn't seem to offer automatic suggestions anymore, so I don't believe you can recreate this). I removed my sim card and unplugged my wifi access point so
        • by Ksevio ( 865461 )
          Sure you can run speech recognition fairly easily on slow hardware (even a raspberryPi). The big advantages for running on a server is you can keep updating the model with new data and do all sorts of fun lookups. The models generally are very large so transferring those would use a lot of bandwidth and take up a lot of local storage
      • Comment removed based on user account deletion
    • by goombah99 ( 560566 ) on Friday September 06, 2019 @11:26AM (#59165740)

      The phones can listen. And they may listen. But they don't have to listen. Two phones does not a sample make.

    • by AmiMoJo ( 196126 )

      The wake up phrase is handled locally by dedicated low power silicon that only recognises one or two phrases. If it was actively recording and decoding speech the whole time your battery would last a few hours. The low power chips don't understand words at all, they just look for a certain sequence of tones.

      Also, at least on Android you can disable "OK Google", or configure it to only be active when the phone is unlocked. I suppose if you were paranoid you wouldn't trust it, but given that all the speech re

    • Re: (Score:2, Interesting)

      I agree, this is TOTAL bullshit. You can take note of all adverts that you receive. Do any of them have anything to do with horses, trains, or mountain biking? If not, just start having discussions about any of them, and see if it changes what adverts you receive. I've done this, and it's certainly a thing.

    • Yeah, but it is not listening _secretly_.
    • When you say "Ok Google/Siri" and the phone "wakes up" and responds.. it's frigging listening - always- period.

      There's a difference between the phone listening, and Google listening. Turn on airplane mode and say "Ok Google". Notice how it still works? Though I don't blame you, the summary and the title are talking about two very different things.

  • by the_B0fh ( 208483 ) on Friday September 06, 2019 @11:17AM (#59165696) Homepage

    I'm not sure how to tell you this, but you can't prove a negative...

    At least not this way.

    • by geek ( 5680 )

      Its not a negative when the phones literally listen for commands 24/7. That is not in question.

      The question is to what degree do they listen. I do not know personally but I do know the track record and history of the companies being questioned and they are not good.

  • by PackMan97 ( 244419 ) on Friday September 06, 2019 @11:21AM (#59165712)
    This is exactly what you would expect someone to say if they were actually secretly listening to everything and everyone.
  • Google (Score:5, Interesting)

    by fox171171 ( 1425329 ) on Friday September 06, 2019 @11:21AM (#59165714)

    I recently went to the grocery store. Asked an employee if they had coconut cream pie. There wasn't any at the time. I left and went home.

    Google popped up a question on my phone, asking about the grocery store. "Can you get coconut cream here?"

    • It must be a total coincidence... I personally talk about coconuts about 200 times a day. And it never asked me about coconuts. So it must not be listening. ( I could have done their research for 10$ in 2 minutes see ) Mouwhahaha.. seriously... This study doesnt prove much IMHO. They can easily spot recordings... and who`s stupid enough to listen to dog food advert in loop. The phone probably flagged it as " owned by moron who`s only into dog food like crazy. No need for advert he`s already hooked to X b
    • Re:Google (Score:4, Interesting)

      by Solandri ( 704621 ) on Friday September 06, 2019 @12:51PM (#59166066)
      • A household averages about 1.7 grocery trips per week [statista.com]
      • At 125 million households in the U.S, that's 11 billion trips to the grocery store per year.
      • Figure Google has been running this program where it asks you about places you've visited for about 5 years (that's when I sorta recall it starting). That's 55 billion trips.
      • About 80% of Americans own a smartphone [pewinternet.org], so that's 44 billion trips with a smartphone.
      • Say Google Now (or whatever they're calling it these days) asks 10% of them about their trip. That's 4.4 billion queries.
      • Say 5% of those questions are about if the grocery store carries a particular item. That's 220 million queries about a specific item.
      • A supermarket carries about 30,000 items [fmi.org]. If you figure roughly 10 brands or sizes per item, that's about 3000 categories of items.
      • Figure on average people pick up a dozen items at the grocery store per trip..
      • Figure on average people ask for help finding an item about every 10 trips.
      • So the odds of a query being for an item you asked about is (0.1)*(12/3000) = 0.04%.
      • That means in the last 5 years, (220 million)*(0.04%) = 88,000 people went to the grocery store, asked for help finding an item, and then later got a Google query asking them about that same item.
      • The latest stats I could find say Slashdot gets 3.7 million unique visitors per month. So figure it's around half a million per day. That's 0.16% of the U.S. population.
      • So of the half million Slashdot readers who read this story, just by chance alone, (0.16%)*(88000) = 141 of them will have had the experience you just described in the last 5 years.
      • Figure only 1 in 100 readers are the type who will post a comment.
      • So on average you'd expect 1.4 people per story to post about having this experience at a grocery store. (Although now that I've said this, it may provoke some of the 140 others to post)

      Congratulations. You are that 1 person.

  • And yet... (Score:5, Interesting)

    by Nuitari The Wiz ( 1123889 ) on Friday September 06, 2019 @11:22AM (#59165720)

    A few months ago I was on lunch with a friend and I mentioned how I'm thinking about going to Europe (and this was before I searched for flights etc) and magically 3h later all the shitty car ads on facebook became KLM ads...
    My friend has the facebook app and uses it, I disabled mine on the phone when I got the phone earlier this year.
    Way too spooky of a "coincidence"

    • by bws111 ( 1216812 )

      This is the same reasoning used by people who claim psychics really work. Just ignore (willfully or otherwise) all of the stuff that they got wrong, and focus only on something that may, kinda, be right. It can't possibly be a 'coincidence'. Except it is. How many hundreds or thousands of other conversations have you had that DIDN'T result in ads?

    • Re:And yet... (Score:4, Insightful)

      by hey! ( 33014 ) on Friday September 06, 2019 @11:41AM (#59165792) Homepage Journal

      When data mining makes a plausible hit in identifying an ad target, it can feel spooky.

      A year or two ago the Internet really decided I was just the sort of person who might take up unicycling. Everywhere I went I was confronted with advertisements for unicycles. How to unicycle videos showed up in my search results when I wasn't specifically looking for them. I complained to my son about this, and he said, "But Dad, that's something you would totally do."

      Which of course is what makes it creepy. On the other hand, recently the Internet has decided that I really might want buy refrigerator ice maker water filters. I have no use for such a thing. Such false positive matches happen all the time, we just don't notice because it's irrelevant information to us.

    • by poptix ( 78287 )

      There's the other alternative -- they've been advertising flights to Europe for a while and you've only subconsciously processed them. Now that you've said something about it to a friend you're consciously noticing them.

      • There's the other alternative -- they've been advertising flights to Europe for a while and you've only subconsciously processed them. Now that you've said something about it to a friend you're consciously noticing them.

        No, I was heavily being spammed SUV ads for almost a full month before that (I drive a Leaf so...). And it was during a period of time where facebook was managing to evade adblock plus, so the ads were particularly annoying and in my face...

    • by ljw1004 ( 764174 )

      A few months ago I was on lunch with a friend and I mentioned how I'm thinking about going to Europe (and this was before I searched for flights etc) and magically 3h later all the shitty car ads on facebook became KLM ads... My friend has the facebook app and uses it, I disabled mine on the phone when I got the phone earlier this year. Way too spooky of a "coincidence"

      I think the most likely explanation is that you're just more predictable than you think.

      I guess you mean that you use facebook via its website either on your phone or on your desktop? Did you do any searches at all related to Europe prior to the ads showing? Did your *friend* do searches for Europe after you mentioned it to them? It'd be easy for a tech giant to see that you're friends with someone, and they searched for Europe, so it's going to show you ads for Europe-related stuff too.

      • I think the most likely explanation is that you're just more predictable than you think.

        Perhaps, on the other hand it was the first time I mentioned going to Europe in about 3 years to someone else...

        I guess you mean that you use facebook via its website either on your phone or on your desktop? Did you do any searches at all related to Europe prior to the ads showing? Did your *friend* do searches for Europe after you mentioned it to them? It'd be easy for a tech giant to see that you're friends with someone, and they searched for Europe, so it's going to show you ads for Europe-related stuff too.

        No I did not search for Europe before the ads showed, and I only use my facebook account on desktop computers. With Firefox using the facebook container. I've gone through the trouble of disabling it from my phone, and also remove Bixby. I don't know if my friend searched for Europe, didn't think of asking at the time.

    • Comment removed based on user account deletion
  • For 30 minutes they played the sound of cat and dog food adverts on loop.

    Google is smart enough to identify a song by listening to 30 seconds of it, even against the backdrop of people talking, bar patrons, etc. Content-matching algorithms regularly flag copyrighted material based on small snippets of it uploaded to YouTube. So I don't think it's tinfoil-hat territory to believe that any app that could listen in for clues on interests, etc. could also differentiate between a repeated advert and regular human speech.

    There's also the fact that one study, conducted on a couple of

    • by Pascoea ( 968200 ) on Friday September 06, 2019 @11:37AM (#59165780)

      Some flaws? Sample size of 2? Using specific words played on a loop? Who "engineered" this study? Were the intentionally trying to not re-create actual use cases? Because they did a damn good job of it.

      You could show me 1000 "studies" that show the phones aren't listening, and I still won't believe you. FAR too much anecdotal evidence and personal experience show that they are. Ironically, one of the cases of it happening to me was related to a dog food conversation with my mother-in-law.

      • It should be pretty easy to prove by sniffling/logging all traffic coming from the phone. I can't take any study about this seriously if they didn't analyze every packet leaving the device especially since it would be trivial to do.

      • You could show me 1000 "studies" that show the phones aren't listening, and I still won't believe you. FAR too much anecdotal evidence and personal experience

        Exactly! A sample size of 1000 is nothing compared with personal experience and anecdotes. In fact, science, who needs it!?

        • by Pascoea ( 968200 )
          The first thing they teach when talking about proofs and theorems: It doesn't matter how many times your experiment yields the same answers, one "wrong" one still disproves the theory. I understand that examples like this [vice.com] aren't scientific studies, but there are hundreds of such examples out there. Those, along with my personal observations, are enough to invalidate their "proof" that phone's aren't listening. I would willingly bet money that in the near future someone smarter than I am will reveal damni
          • by Ksevio ( 865461 )
            Ok, maybe show one study that they are listening? You weren't taught very well if you weren't taught to look for alternatives for the 1 in 1 thousand times you saw a different result
        • by cirby ( 2599 )

          This was a sample size of TWO. Over a very, very short period.

          Not a thousand.

          If I wanted to set up a phone "listening" service and didn't want it to be easily found, I'd certainly set it up so it would dodge a test as simplistic as this one.

          For one thing, make sure it didn't engage on a new phone, since testing labs will always use "clean" setups. Only use phones that are well-used, with a good history of miscellaneous net traffic. Those are your high-value "networked" customers in the first place.

          For anoth

    • Re: (Score:3, Insightful)

      by hey! ( 33014 )

      The study also assumes that if the app vendors are listening to what we say, they use the most simplistic thing imaginable with that data. "Oh, somebody within earshot said the world 'dog food', let's start spamming the user with dog food ads."

      If these guys are listening, they aren't listening to trigger discrete events; they're in the business of building a profile of your behavior and habits.

  • It would seem like an easy thing to test for - leave various apps running and talk about something new - so where are all the studies?
  • by Alain Williams ( 2972 ) <addw@phcomp.co.uk> on Friday September 06, 2019 @11:35AM (#59165772) Homepage

    Google, Facebook, NSA, GCHQ, Israeli Spyware Firm NSO Group [slashdot.org], or someone else ?

  • by Anonymous Coward
    Have we learned nothing from Snowden?
  • by chuckugly ( 2030942 ) on Friday September 06, 2019 @11:39AM (#59165784)
    I'm not completely convinced that playing the ads is the same as conversation about, for instance, dogs. *If* "they" were listening they could probably distinguish between those two very different types of audio. If the machine can't, surely the developing nation contractors they hire to do the real heavy lifting could do so. I'll file this under "right for the wrong reasons" tab.
  • My wife and I have discussed things recently only to have those topics reflected soon after in her Facebook feed. Unique topics we had not discussed before. That we had not searched for. New and unique. Travel plans were the most common ones, but others included legal issues, and specific medical issues not previously discussed by us because, well, they actually didn't affect us personally. Undeniable.

    I haven't noticed this in my Facebook feed (She's got an iPhone, me, Pixel 3a), but I wasn't observant befo

  • One of the arguments that phones are listening is that sometimes, people will have a conversation about something, and suddenly, without making any social media posts or sending messages or emails about the topic, ads will show up for that thing.

    For example, I was discussing the fact that I hadn't had a steak dinner in some time not too long ago. Interestingly, ads for a local steak house showed up.

    How did they know? Well, some folks think the phones are listening. This is 1984 scary. But the other poss

  • by TexasDiaz ( 4256139 ) on Friday September 06, 2019 @12:06PM (#59165892)
    Oh please, these technology researchers clearly don't know much about technology. Let me give you an example. I have been watching a spanish TV show on Telemundo called Exatlon, but since I don't speak Spanish it's quite difficult. So, I tried using Google Translate on my phone to translate some of the audio - however, Google Translate knew that the audio was coming from a device and wasn't "someone speaking", so while it showed that it was receiving audio it didn't perform any translations. However, if I then repeated what I just heard, it immediately picked up and translated that. (I can only believe it determined that based on frequencies it was or wasn't hearing from the audio coming out of my TV vs what frequencies it was hearing from me). If the researchers just played TV ads, any technology using the same system Google was using would be able to ignore everything it heard from a device, and only was trying to listen to "real" people in the room. By the way, Amazon Alexa sucks, it still responds to the wake word when it hears it from a TV, whereas Google Home doesn't respond to the wake word from a TV.
  • The phones have GPS, so the apps know when they're in a testing facility and turn off the surveillance -- duh. They also have various sensors that tell them they're not being handled ... /tin-foil-hat

    Sounds far-fetched, but VW did this to pass emissions tests -- the vehicle detected when the vehicle was being tested and enabled emission controls during the tests. It did this by (from How They Did It: An Analysis of Emission Defeat Devices in Modern Automobiles [ucsd.edu]

    The defeat device in Volkswagen vehicles used environmental parameters, including time and distance traveled, to detect a standard emissions test cycle: if the engine control unit determined that the vehicle was not under test, it would disable certain emission control measures, in some cases leading the vehicle to emit up to 40 times the allowed nitrogen oxides.

    Defeat devices like Volkswagen’s are possible because of how regulatory agencies test vehicles for compliance before they can be offered for sale. In most jurisdictions, including the US and Europe, emissions tests are performed on a chassis dynamometer, a fixture that holds the vehicle in place while allowing its tires to rotate freely.

  • Here is the name of a company that produces software used to allow app producers to listen in.

    The apps use software from Alphonso, a start-up that collects TV-viewing data for advertisers. Using a smartphone’s microphone, Alphonso’s software can detail what people watch by identifying audio signals in TV ads and shows, sometimes even matching that information with the places people visit and the movies they see.

    Quote from here: https://www.nytimes.com/2017/1... [nytimes.com]

  • Well sure I can hear everything, but I promise not to listen.

  • I've had this same conversation with non-technical family and friends. It seems to be a combination of not understanding networking as well as just a general mistrust. The amount of data that would be generated if _every single phone_ was constantly streaming data to their home base would be staggering. Not to mention multiple apps, streaming all voice data, to multiple home bases (Google, Apple, Facebook, Amazon) would cripple your phone's ability to do anything. As well as a family all connecting to t
    • Although you are correct, it doesn't have to be a constant stream. Maybe it only transmits when it detects speech at a certain volume, or certain keywords. The criteria can change on the whim of the corporate overlord. We do know that phones constantly listen: that is how the whole "wake word" that the various agents (Siri, etc) works. The study itself is idiotic, because even if the phone isn't listening now, it could 5 minutes later after. It runs a closed ecosystem, so the behavior can change minute by m

  • I'm calling BS on this.

    For starters, this is not valid. Taking some unused phones and subjecting them to 'test condition' audio seems completely bogus. One of the things these phones do, over time, is 'learn' about their owner.

    Feeding the phone 'test condition' audio with no actual normal use... invalid.

    Second, even if that was valid, and it's not... who's to say the operators of the backend don't know what's going on and invalidate the test by making the phone respond differently than if it wasn't in a l

  • Google Maps did it to me. Picked up a conversation in the car and mirrored the information in my Youtube recommendations.

    • by Ksevio ( 865461 )
      Do you use voice commands to control it? If not, check if it has the microphone permission and you can disable it if it does. Then you can be paranoid about other apps if it happens again
  • It does seem like they're listening, doesn't it?

    One of the leaks from Facebook said that the problem isn't that you're being recorded, but that their machine-learning algorithms pattern match to figure out what you're likely to think of next based on their big data.

    Another commenter here said that he asked the clerk at a store for some coconut cream pie and then he got an ad for one when he went home. So there are four possibilities:

    1) they were listening.
    2) they pattern matched and failed to predict your

  • When people will just willingly give you all the freaking data, and you can use it with NNs to generate even more data?

    • That is probably more accurate. It isn't worth the effort. They already know what location/age/gender/income/etc you are, which is enough to market to you.

      • by Z80a ( 971949 )

        They also have the way you write, all your pictures, the way you talk, everything you searched, all your social connections...

  • I'm pretty sure I read an article a while ago where they interviewed a developer who wrote some of this "listening" software for social media apps. And his explanation was more that apps like Facebook can and do/did listen occasionally, but only for certain "target" keywords.

    Basically, they have no interest in what you're saying until you say just the right word that they've been paid by someone to drive targeted marketing based on it.

    If a Japanese airline company pays for ad placement, then for the length

  • I am skeptical, speaking as an ancient phone hacker who knows that the Infinity Transmitter dates back to 1973 when it came out of the NSA. I do know that Covington of Wandera (originator of this "study") was with Cisco, and Cisco should now be considered a Chinese Communist Party company since they forked over their source code to the CCP to do business there.
  • Comment removed based on user account deletion
  • Missing the point (Score:5, Insightful)

    by MobyDisk ( 75490 ) on Friday September 06, 2019 @01:46PM (#59166288) Homepage

    If you wanted to know what the Facebook app is sending to Facebook, then modify the phone to track what is being transferred. Or have we reached a point where end-to-end security is so good that we can't even read the communications send by our own devices?

    We should be less concerned about what is being sent, and more concerned about the fact that we don't have an easy way to find it out.

  • Oh gee they used a whole four phones for their 'study'/'experiment'/whatever you want to call it? LOL, their 'findings' are meaningless.
    Conduct a proper study: have thousands of people volunteer for the 'study'/'experiment', select thousands at random, conduct a background check on them to be sure they're not Google/Facebook/whoever plants intent on poisoning the data with false negatives, have them all sign NDAs so no one but the researchers and the participants know who is actually involved, and spend a
  • They know where you go well enough, they can easily know when to listen for keywords and when to wait to send it so it doesn't impact your data. It can tell that nothing is going on, so doesn't activate any reporting. It also could be recognizing that the data it's getting a stream of is advertisements, not conversation, so it just stops processing it until a significant change. Or perhaps the reporting waits for WiFi connection anyway so as to not use up someone's data, so it seems no data is sent during t

  • I conducted a similar experiment a year ago. I wandered around my apartment which contains multiple phones, computers, and microphones talking for days about how much I wanted to buy camping equipment: Tents, sleeping bags, a canoe, you know... camping stuff.

    I waited for the ads to show up in my browsing experience, but....nothing.

    Then again, that was a year ago...

  • by uncoveror ( 570620 ) on Friday September 06, 2019 @01:55PM (#59166330) Homepage
    Even if our phones aren't spying on , our TVs are! http://uncoveror.com/vchip.htm [uncoveror.com] And ceiling fans, too. http://uncoveror.com/fans.htm [uncoveror.com]
  • What did this prove? (Score:4, Interesting)

    by chiefcrash ( 1315009 ) on Friday September 06, 2019 @02:03PM (#59166372)
    What exactly did this experiment prove?

    The claim was that phones were listening to people's conversations, so they tested this by using audio from advertisements? Didn't an advertisement that activated people's smart speakers cause the industry to start testing for and filtering out non-spoken audio like this? What sort of audio does the advertisement consist of?

    The article says the researchers noted "no significant spike in data or battery usage.", but then also notes "interestingly, the study found that most of the Android phone apps seem to consume significantly more data in the silent rooms with many iOS apps using more in the audio-filled rooms." So their own data shows the devices act differently in silent rooms vs non-silent rooms. The analysts say they aren't sure why this is, but we can assume it's not phones spying?

    The company co-founder is "confident that the overall results show that any secret transfer of significant data is not happening" after such a limited test? The experiment only ran for 3 days!. What if it stores and forwards data on a weekly schedule?
  • Always assumed if this were occurring it would not be Facebook doing it directly rather a part of their third party/network/API for in app advertising.

    You install some third party app... it does the microphone spying then feeds the context to Facebooks ad network.

    If it were me making a serious attempt at detecting this I would start with app lists of all those involved with creepy "coincidences" I could find and do my testing from there.. not with a clean slate but with app lists in the wild as devices are

  • Driving with google maps open. Began talking with my wife about rentals in Italy, where we were going the following week. Did not say "ok google" or anything approaching. Maps begins helpfully popping up bed & breakfast locations around us.

  • by twocows ( 1216842 ) on Friday September 06, 2019 @02:50PM (#59166624)
    Oh hey, it's Wandera. We use them for tracking our employees' cell data usage on their company provided phones so when our report goes out of people using too much data and they come to us and tell us we're lying or the system is lying, we can point to our records that show they used 38GB watching ESPN videos. They still don't admit wrongdoing but their data usage magically drops the next month in almost every case.

    Good product but I question the result. Not even a month ago, I watched a YouTube video with millions of views where a guy showed the ads he was getting at the start of the video (sans dog toys), then talked about dog toys next to his phone for like 3-4 minutes, then showed his ads at the end of the video which were about dog toys over several different websites. Maybe that's not proof, but it's suspicious as hell.

    I can come up with a number of reasons they might have reached this result. Perhaps Google implemented counter-measures for experiments like this. It's probably not hard for an algorithm to tell the difference between a looping advertisement being played repeatedly and a person organically mentioning a product a couple of times. Perhaps there are other flaws with their methodology exist that I didn't think of. Or perhaps a company that puts out software to track data usage has a compelling interest in making people less paranoid about companies tracking what they're doing or saying.
  • by ShieldW0lf ( 601553 ) on Friday September 06, 2019 @03:34PM (#59166864) Journal

    I can't comment about listening to conversations, but advertisers are using phones, computers and televisions to surreptitiously listen to the environment as a way to deliver targeted ads, and have been for a long time. The technology isn't hard to build, I could do it myself if I was so inclined. There's a relevant article from 4 years ago about a company called SilverPush on arstechnica.com.

    https://arstechnica.com/tech-p... [arstechnica.com]

  • by LostMyBeaver ( 1226054 ) on Friday September 06, 2019 @11:30PM (#59168088)
    Holy sheep balls!!!

    This was research? What kind of research? The kind where we wear tin foil hats?

    Get an Android phone and an iPhone.
    Connect them to the Internet via Wifi.
    Use an MDM solution to install invasive enterprise certificates.
    Play sounds and capture all data from phones.
    Use private keys to decrypt traffic
    Find anything which seems suspicious

    Why not download the Android source, instrument it and build it?

    Why not disassemble the facebook app, write a "trace finder", follow the trace back from socket system calls?

    This was shit research

Work without a vision is slavery, Vision without work is a pipe dream, But vision with work is the hope of the world.

Working...