Facebook denies 'listening' to conversations
bbc.com
bbc.com
Facebook is listening to your data--to all of our data, all at once. They have locations, searches, clicks, messaging, photos, hashtags, and any other form of browsing patterns for everyone in your country, everyone in your neighboorhood, everyone in the same room as you.
They have enough data with such advanced analysis that on occasion they can get really close what you're thinking/doing without you explicitly telling them. They're doing this frequently enough to creep a lot of people out and they're only going to get better at it over time.
What we have here is something like Turing test for privacy: a sufficiently advanced amount of data and analysis will be indistinguishable from surveillance.
It's simply unnecessary to listen with a microphone.
Edit: found this slide from F8 in 2015 where they said they store 300PB and process 10PB (with a P) of data per day https://www.instagram.com/p/0tUjrQKH6R
We like to pretend we're unique and powered by free-will but the reality is most things you do are fairly easy to track and predict.
We're not in 1984, but certainly the capability is _already_ there. The difference between us and 1984 is only really one of intent, and you can guess how long that'll last.
I'm not convinced intent isn't there; I'm also confused why'd you assume that. As a free software activist, I wonder have you every checked EFF or FSF websites. I'm not saying EFF and FSF are saving the world (no, they are not) but if you browse what they're dealing with, you can easily see the intent is there. It is simply hard to do all these without people thinking intent is not there, which is why nothing is done as explicitly as 1984. For example, Windows 10 has built-in ads, INSIDE operating system, your OS, which you already paid quite a price, forces you see ads; and tracks your behavior "to optimize user experience". People are not worried about this since ads are such an integral part of our lives. But Microsoft perfectly intents to track which websites I visit, what do I eat, what do I like, where would I go. Maybe you can say intention is "different" as in in 1984 they use my personal information to prosecute me, whereas this is not what Microsoft is after. Well, then this brings us to some Foucoult reading. In a world where all my information collected what is the difference between using all my information for a specific cause, when this cause can change just tomorrow. What if NSA asks Microsoft to give bunch of people's data? (we all know this already happened, but it is better practice to ask questions)
I am a programmer, I use my computer all the time, pretty much every single awake moment of mine. I don't want to be watched. But as you said it is depressingly hard to achieve this, since it feels like only 0.1% of the population shares this same desire. Most people are okay being watched, and this is part of the problem.
I'll skip the futurama reference for now, but for me ads are far rarer than in the 90s. Adblock handles most things online, iPlayer, Netflix and amazon don't really have adverts on TV (they have a graphical menu but I would call that an advert), Spotify doesn't have adverts. Sure there are a few adverts if I go to somewhere like london, and Sainsbury's advertise their own deals in store, and of course products on the shelf are adverts in of themselves, but the only advert I have seen in the last 24 hours was one raising awareness of cistic fibrosis in a billboard outside of Sainsbury's. Oh and the train announcement saying "the on board shop is open for teas coffees and light refreshments". There wouldn't have been any product placement in the episode of once upon a time I watched last night either.
The TV jingles I remember from childhood are still etched in my brain, the adverts I see in this decade are ones I hunt out depberatly (the John Lewis Xmas advert for example, or film trailers, or something about amazon delivery people stealing your sofa that was recommended at work)
I did see adverts on Saturday night when we went to see Thor. I don't recall what was advertised other than odeon limitless, the trailers are adverts too of course, alas they don't really work - we'd rather see more films at the cinema than we do now, but getting babysitting is a pain. We missed Spider-Man and kingsmen for example.
If anything there are more opportunities to avoid adverts than ever before.
I am not sure I agree with that. I think the vast majority of people react to the same way to surveillance, they at most don't like it, or outright hate it.
But most people are not conscious of this surveillance. If there was a guy following them absolutely everywhere, all day long, and logging on a notepad everything they do and everything they say, they would go absolutely mad. It's just that current surveillance, through cookies, CCTVs, server side logging, is too stealth for people to notice.
Implies only one can be right. There doesn't have to be only one answer.
Some say the world will end in fire,
Some say in ice.
From what I’ve tasted of desire
I hold with those who favor fire.
But if it had to perish twice,
I think I know enough of hate
To say that for destruction ice
Is also great
And would suffice.
I don't know which services provide this or how they gather data, but the thought of anyone being able to download a detailed file on you with only an email or phone number is scary.
Think about it: tomorrow is Monday. How many folks on this board will be working on analytics and tracking once they get in to the office? A lot of us, guaranteed. And how many of us will take a principled stand, swearing not to work on products until surveillance features are removed? A few of us, guaranteed.
We're living in both 1984 and Brave New World.
See this illustration between 1984 and Brave New World.
Even if they just knew the price of the purchase, combined with the job (burn risk) and location data (traveled from work to nearest store with pharmacy during the day), it's possible that you might be able to infer a burn with enough accuracy to be valuable to advertisers.
It's not possible, it's simply true. In fact, this feature of the facebook ad platform is available to any advertiser on Facebook/Instagram/etc, whether you have $100 to spend or $100M.
You simply upload your customer's purchases to Facebook with data such as zip code, email, etc etc which Facebook then uses to optimize your advertising budget and find more customers similar to the ones you already have (the feature is called look-alike audiences).
source: http://www.bbc.com/news/technology-40027706
If Google has access, I am sure Facebook does it.
They have billions of users, billions of ads, and trillions of pieces of metadata. Sifting through that to produce a guess like "User 1234 burned themselves and would be interested in product 5678" would be an amazing and scary piece of AI.
Not to mention that Facebook are limited by the ads they can sell. The manufacturer of the burn cream doesn't give Facebook a bucket of money and free reign. They choose how and where the ads are shown. There is currently no option for "Users who recently burned themselves".
From the links that m_ke posted, facebook already has ties to loyalty card information, so it's very possible that they didn't need to do any inference, they just had the data directly.
In any case, the main point is that facebook doesn't need to listen to what people are saying, it has a ton of other data streams that could explain the stories in the link.
I’ve had an account since 2005, even if I use it much less than I used to, you’d think they’d have enough data to do better than that.
Building this kind of predictive models is not black magic. This is what companies do to figure out who they should target on direct marketing campaigns. Applying some magical deep learning dust can quite likely improve the results (in Facebook scale), but it's not mandatory. Computationally the hardest part is building the models. Once they are done, applying them to millions and millions of customers is more straight forward.
When considering what is possible, you need to also consider the stakes. Facebook ad revenue is something like $10 billion per quarter[1]. To make more money, I can think of three ways: add more users, make users spend more time looking at their feed or make more money per ad slot. Better targeting means more money per ad slot since some customers are paying directly based on the clicks. Since improvements lead to better revenue, extra hardware costs are easy to justify.
[1] http://www.adweek.com/digital/facebook-raked-in-9-16-billion...
P(product_class=bandage | job=factory_worker, pharmacy_visit_last_month=1) >> P(product_class=bandage)
There's a reason they got so dominant over here - they were simply way ahead of the game on offering what their customers actually wanted, because they were the only people who actually had a clue.
That has of course changed now.
Coincidentally, I recently had a friend convey a story where the Facebook app suddenly recommended a new connection immediately after certain information was spoken in a verbal conversation. This instance was particularly damning because, due to the sensitivity of the information that was spoken, it had been very carefully kept out of the digital footprint.
Once Facebook recommended me to connect to guy with an interesting name, and I wondered where I saw this name before. I looked up my emails, and I saw that I bought something from him on eBay several years earlier. I know I never gave Facebook access to my mail account, but guess what, I'm pretty sure the guy gave access to his. Facebook saw that we sent mail to each other, and it asked me if I wanted to be FB friends with this random eBayer...
I had this happen personally. Had a conversation about my work with this girl I know which ended up being mostly about project management. The girl told me later she started getting ads for project management stuff.
It's not a subject she's interested in, has ever searched for information on, and has no relevance to her job working the counter at a sandwich shop. We have no social media accounts in common and don't even talk that often.
Sorry to 'actually' you but that's phrased like Clarke's third law: "Any sufficiently advanced technology is indistinguishable from magic."
However, as you said, they don't need to. Localization and usage metrics alone can tell you an incredible amount of extremely detailed information.
For foreground, I suppose it'd be fairly trivial to tell if the app was making calls to AVAudioRecorder.
Unfortunately, I don't think they would. They'd get a slap on the wrist fine, maybe have one or two senior executives resign, they'd post a mea culpa, saying how they now it was wrong and promise to do better, and then everyone would forget about it a few weeks later when some new thing started to dominate the news cycle.
It's a pretty easy test - they just said "cat food" a bunch of times and ads for cat food came up in their feed the next day. I'm sure if you were interested you could probably do the same experiment with some other product you've never posted about or talked about online and come up with similar results. It seems strange that you would just assume that the people who are bringing up this issue are simply misinformed, or do not also know that facebook also looks at other parts of their data. I think most lay people understand how data mining is leveraged these days.
I'd like to point out that this video doesn't actually provide any evidence that Facebook was listening.
They said they talked about cat food all day, then they showed an ad on Facebook for cat food.
There's no way to verify that they didn't make the video of the cat food ad on Facebook and then talk about cat food all day.
Nor is there any evidence that they talked about cat food all day at all. No recording of the conversations etc.
In fact I'm a little confused about how this is even a question.
I'm not a programmer so I don't know, but: Wouldn't it be trivial for someone who knows how to write Android software to monitor if an an application is accessing the audio input device?
I mean, I know that on Linux you can monitor whether or not a device is being opened.
Why doesn't someone check if Facebook is accessing the audio input?
As soon as you start assuming that the OS is giving the facebook app access to the microphone regardless of whether you allowed it in the settings, things start to get absurd
Or even by MITMing the connection and looking at the network packets. But yes, it's not outside the realms of possibility to hook into the microphone driver on a rooted Android phone and check when it's being activated.
I'm not one to leap to Facebook's defence and if this is happening it needs to be shut down ASAP, but I suspect that there would be at least some credible evidence out there if it were indeed the case.
I feel like there'd be hundreds of these videos if it was real. How many people tried to make this video and it didn't work? I hope someone buys 10 new phones, opens 10 new Facebook accounts, and does some statistically rigorous testing.
To be a scientific test it needs a list of products, and then randomly assign half into a study group and half in to a control group. Then talk all the time about stuff in the study group but not the control group. And see if there is a statistically significant difference between the ads shown for products in the study group vs the control group.
Explaining away doesn't explain. For an average, normal FB user, this was as scientific as they could get. It would be nice to see someone repeat the experiment with a pro-hacker type on hand and a packet sniffer ++.
There are too many people giving a company like fb, who have form on the implementation of morally and ethically dubious practices, the benefit of the doubt, all the while dismissing any and all claims of users. It looks fishy at best, and I'm not name calling, it isn't allowed on this forum, but there are a disproportionate number of FB defenders popping up wherever this story surfaces.
So is Google. What makes FB even creepier is that they don't just watch everything people do, but they manipulate people in controlled experiments to extract as much money from them as possible.
You just described A/B testing. This is a technique used by every successful company. The only thing unique about Facebook is the data.
Definitely. Basically every major website has facebook code on it for social sharing and for "analytics". Facebook surely uses this analytics data to track you from site to site and know exactly what you're looking at. If your browser is signed into facebook, they know exactly who you are too.
There's an issue of whether facebook is, despite what they say, doing some sort of data analysis that's relevant to advertising, in addition to all the other listening and analyzing they do with the rest of our data.
What’s surveillance if not “sufficently advanced amount of data”? They’re indistinguishable becaus they are the same.
Facebook isn’t “collecting data”. It really is following you.
If I had to give my father some reasons why this should concern him, I'd come up blank. Maybe you can fill in a few.
https://www.wsj.com/articles/chinas-new-tool-for-social-cont...
That's why employers requiring social media information is illegal in a dozen states or so. (Illinois, I know for sure. There are others.)
Once I realized that, it was hard to think of reasons why their actions are wrong.
Facebook has data sharing agreements with many, many companies. It's most likely that the cream was bought with a card then linked together with your facebook profile. It's not that hard, especially when you use the same email identifier for all those loyalty cards you have.
EDIT
A video that explains what Facebook does https://www.ted.com/talks/zeynep_tufekci_we_re_building_a_dy...
I figured that youtube got the credit card info from the purchase but I was shocked at how quickly that translated into something showing on the screen. I've always thought about these things in the back of my mind but this was still eye opening for me.
I'd never had it happen to me personally since I aggressively block all ads :)
Was the pharmacy a national chain, regional, local, or one-off?
In my experience in data exchanges with other large companies, data is usually shared and processed in batches where even an hour lag would be considered "very quick".
The funny thing is, this is a horrible use of ad data. With a few exceptions (luxury cars being one, making someone feel good about their purchase) showing ads for a product someone just bought is the height of uselessness.
Facebook and YouTube are selling ads to people who have already bought their product.
So if I pay for petrol with a card and also buy donuts and coffee the bank and the shop cannot market to me according to what I bought, but if I join a loyalty card scheme then I'm opting in and the shop can, but not the bank.
In Norway btw, where govt. are strict about this sort of thing.
It sounds like someone is trying to game the advertiser by correlating ad views with sales.
Shhh. If advertisers realize how utterly wasted most of their advertising spending is you'll destroy about 80% of "tech" companies and plunge the world into a global recession.
(Is it really easier to believe that we, the "smart techies" of HN, just know so much more and can see so much more clearly than everyone else?)
What ways do you think they company might be able to use that information to sway the outcome of the election (passively or actively), and do you think there is a greater than zero chance they would try to do this?
Personally, I am so used to corp speak and word play by major companies that it is difficult to take anything at face value. Therefore, I think the nature of the question asked to Facebook should be different. It should be on what they do rather do they do X. So, here's my question - "What does Facebook use data gathered from user microphones across any devices for?"
Edit: https://www.theverge.com/2016/6/3/11854860/facebook-smartpho...
In a statement issued on June 2nd [2016], Facebook said it "does not use your
phone’s microphone to inform ads or to change what you see in News Feed."
Original comment:
If they had said something like "We don't use microphones to gather data to drive personalisation on users profiles" there would be comments here on HN complaining about corporate speak and how they didnt just flat out deny using it for advertising.
Apps get busted for doing dodgy things all the time. The Facebook app has been decompiled and network traffic inspected and studied so often that if this actually was true then it would have been proven by now. Instead, all we have is this weird 'survivorship bias' where one time someone said something out loud there was a tangentially related ad on FB.
Today I was looking up bars in Budapest and found one on Google. I started looking at images, at the only one that loaded so far on the slow internet was of food. "Wow, Google knows me so well. All I care about right now is food and the first image it loads is of food" I said out loud to my friend. Eventually the internet caught up and the rest of the images that loaded were of food. It only showed me food photos (in a bar thats more known for its decoration and alcohol). Nothing more than a coincidence.
"What does Facebook use data gathered from user microphones across any devices for?"
That particular question, in context of these allegations, would presuppose that they use microphone data for something other than what the user clearly is aware of, e.g. voice chat.
So an answer like "to provide users the service they need" would undeniably be spun by media to be: Facebook uses microphone to tweak news feed. The longer the answer is, the more likely it can be misunderstood and taken out of context.
You shouldn't expect a public company to answer questions like that, simply because the media will gladly take the answer out of context. You need to phrase it specifically, so they can answer without being misunderstood.
Edit: Actually, FB will ask for microphone permission if try to record a video. Based on every other app that records video also prompts for permission to record using the microphone, I'm lead to believe the microphone permission is required on iOS if you want to record video.
FB Messenger does not prompt for Microphone until you record a video or voice memo.
Amend their privacy policy.
>That's pretty fucking clear to me.
Said nobody who understands law, language, or politics. It's absolutely fucking clear to everyone who reads with discrimination what this means. It means they are using your microphone to vacuum up your information. That they added the qualifier "..to inform ads or to change what you see in the News Feed" gives them the verbal and legal wiggle room they need to do absolutely anything they want with your information. For example, they could use your microphone to monitor, record and quantify everything they hear to "adjust a user profile" that they keep on all users. In that case, they could simply claim that the ads you see and the news feed you get are based on your user profile.
>How could they respond to remove all doubt they're listening?
They could very simply. "We do not listen to your microphone or record content, data, or metadata from your microphone in any circumstance". Without qualifiers of any kind.
In this case, when I read a sentence like "We don't use your microphone for ads" I end up doing that game where you re-read the sentence over and over and emphasize a different word each time, trying to guess which one is the magic word.
The risk is incredibly high, and the payoff not that great.
What we're seeing is simply that people fail to understand how much data they leak, and how effective Facebook is at putting it all together. They assume it must be magic, listening to their words, listening to their thoughts, but the reality is they don't need to, because we are _voluntarily_ telling them all this info already.
Somehow this kind of phrasing turns several lights bright on my bullshit detector.
I guess another possible explanation is some very sophisticated combination of location tracking and Kellogg’s targeting people they know will be primed by that promotional item. And then of course it could always just be an amazing coincidence.
All I know is I revoked microphone permission after that.
EDIT: Something just occurred to me. Facebook doesn’t have to be the one using your microphone to target ads. The ad buyers just have to have a way to match your phone to your ID (which they can easily; there was a site posted here a while back that sells that) and then listen in through ANY one of the apps you have installed. If this were the scheme, Facebook could reasonably get away with saying “WE don’t use your microphone to target ads.”
More on-topic: I recall reading that the Facebook app would activate the microphone when one was posting a status update, apparently listening for things like a movie playing in the background and such.
ETA: Related Snopes articles:
https://www.snopes.com/computer/facebook/facebooklisten.asp
https://www.snopes.com/2016/06/04/professor-claims-facebook-...
The "People you may know" anecdote below is, similarly, some combination of location, wifi sniffing, luck, and confirmation bias.
If there was some third party ad library that was installed across multiple apps that streamed audio and/or performed local speech recognition on everything you said don't you think somebody would have a) noticed by now, b) a security researcher figured it out, or c) everyone's battery's would be running red hot all of the time.
Teams at Apple and Google reverse engineer and debug the top 100 apps on their platforms all the time (there are teams dedicated to doing it). Facebook is also not the NSA - people leak stuff from them all the time. You're suggesting that either everyone is complicit in this, or there's a conspiracy so large and technically complex that it's managed to be kept secret all this time with no technical evidence outside of anecdotes.
I’m not sure I trust that security researchers can be relied upon to catch everything. There are so many flagrant abuses of user privacy that go unchecked or remarked upon and nevertheless ignored that such an avenue of prevention seems highly unreliable.
Also you’d be surprised at what people are willing to do when you define a certain culture and incentive structure. While you may dismiss it, I find it frankly astounding that the NSA program managed to metastasize into the behemoth it became before it was finally outed to the public. And even a cursory reading of behavioral psychology will help to explain how that can happen.
I also know that Facebook perversely prioritizes metrics above pretty much all else. And it means they’re perfectly willing to indulge in dark patterns if the numbers line up.
Once they have a business model that does not revolve around selling data about their users to their customers, claiming they don't spy can be something other than a bad punchline.
Anyway, I want a LED on my camera and microphone. And a hardware switch.
now, I also remember them writing that they only listen when the app is open and a status is being written, but I'm not sure about that statement. Anyhow, the key thing is that Facebook does overtly and publicly use your microphone with its app.
I suspect Facebook is falling in that same boat - technically following rules so they can answer questions the right way. But we're not seeing the smirks on their faces as they give that answer.
I was in the car with my friend once (both our phones were in the car with us too), and we were driving and it started raining. His wiper blades were shoddy and doing a bad job of clearing his windscreen. We had a 5 minute discussion on cleaning blades, where was the best place to get a replacement, etc. Neither of us picked up our phones, or did a search on 'wiper blades', nor asked Siri etc. about nearest places to buy.
But the next day, for the first time ever, I started seeing ads for wiper blades on my feed. They disappeared after a few days, but it was spooky, as it was for an item that I have never ever discussed, or seen on my feed before, or since.
Another anecdote. I was sitting working in my office the other day, when my son came in to talk to me about something. While chatting, he picked up a bass guitar that sits behind my desk and started playing the riff from 'Seinfeld'. I may have said something like "Oh, the Seinfeld theme - you know that was played on a synth and not an actual bass guitar?". Our conversation though, was about something else altogether.
A day later, on my feed, I see an ad about "How they played the Seinfeld theme" [0]. A totally specific ad like that, within 24 hours, totally unrelated to any searched I had done in my past. This one too disappeared within a couple of days.
Could be something like this: Facebook knows from location data you were driving/riding together. Soon after your ride, your friend searches for wiper blades. Facebook then shows you wiper blade ads based on that knowledge.
1. It is just as unscientific as saying that Facebook MUST be listening, without finding any evidence in the workings of phones, app source code, etc. Almost by definition, there is no experimental evidence for the Baader-Meinhof Phenomenon.
2. It posits that everyone who has experienced this is unsophisticated and ignorant about cognitive distortion. Clearly, some people who believe that Facebook is listening to them talk also believe that there were no planes on September 11th, etc. However, many many people today have a good understanding of psychology, are well aware of (basic) cognitive biases, have thought about these experiences. People also have experiences of unlikely events but which have well understood or easy-to-estimate probabilities, which didn't occur much historically. Lots of people can distinguish quite well between their own illusory biases and events which are coincidental, but not 'in the mind'. Some of these people seem to also believe their speech has found its way into advertisers' profiles.
Since then, my feed has been a bland sea of adverts for mattresses, knives etc. that almost all of my friends complain about seeing.
I don't use ad blockers, due to my development work in web apps etc. I like my browsers to be as stock as possible, and I do actually at least glance at nearly all ads I see in FB etc. just to see how their targeting is being skewed by my search history or messaging online.
These were two quite distinct cases of ads that I've never seen before, which popped up within 24 hours of me talking about (but not doing anything else online related to them) and then disappearing soon after.
Sure, an anecdotal sample size of 2 could just be pure coincidence, but something about it still doesn't sit right with me.
The link about friends search history and close GPS co-ordinates also makes sense.
Later on when I checked my phone I started getting ads in the Facebook feed for windscreen repairs.
> Here’s how it works: if you choose to turn the feature on, when you write a status update, the app converts any sound into an audio fingerprint on your phone. This fingerprint is sent to our servers to try and match it against our database of audio and TV fingerprints. By design, we do not store fingerprints from your device for any amount of time. And in any event, the fingerprints can’t be reversed into the original audio because they don’t contain enough information.
And my previous comment :[2]
> Notice how they say they do not store sound or recordings. But what if it is converted to text?
Notice below they say that the fingerprint cannot be reversed into the original audio because it doesn't contain enough information. Text cannot be converted to the ORIGINAL audio but it can be converted to audio.
This may be just me being paranoid, but I do know that legalese is designed exactly to say something without saying everything.
[1] http://newsroom.fb.com/news/2014/05/a-new-optional-way-to-sh...
If it’s syllable- or word-length fingerprints it’s not much solace that they won’t be able to get the original audio.
You just forget all the other signal you've giving. Funnily enough, last week somebody was trying to convince me that Facebook was recording their every word.
"I was telling a friend about a recent trip to Thailand", they said, "and I immediately saw an ad for Thai Tea". I asked if he'd used Facebook when he was in Thailand. He had.
Which is more likely: Facebook is recording every conversation of every user on their app and mini them for ads or the advertiser in question just targeted people who had recently been to Thailand.
I mean, really?
So yes, really.
Keep in mind, this was someone who would have been completely new to podcasting and have never searched for podcasts or podcast apps before.
Made both of us highly suspicious.
Now about the confirmation bias thing, it seems more likely to me that your credit card history or the store member card purchase history is being mined.
Someone messaged me on Facebook Messenger, I responded, put my phone back in my pocket, and finished the transaction, and then left Lenscrafters and walked down to the store that I was actually in the mall to see.
Within the day, I started seeing contextual ads for eyeglasses in my facebook feed, along with ads in other apps for eyeglasses.
I'm only suspecting that there's a service that does some flavor of ultrasonic beacon/listening technology, and facebook (along with DFP and the other large ad markets) are buying that service in troves.
The other explanation is Facebook has data sharing agreements with the payment instrument you used for the wipes.
By contextual ads, I mean, "News stream ad/in game ad" -- basically, anything that had been targeted to be sent to my device.
I believe I read something about Facebook using DataLogic info (which tries to tie payment instruments to advertising clients), but, apparently the music supplier for LensCrafters also has "Audio Beacon" technology for their clients to use (http://us.moodmedia.com/social-solutions/), so, it could be one of the two.
This requires infrastructure at a really large scale I don't think even FaceBook can sigh up for that. And if they have a legal department, which I assume they do and a functional one, those ideas would be killed before it moves to implementation.
No evidence that this is what they're doing, though.
So basically just listen for any time someone says a variant of "I need/want x"?
I took advantage of the option to see why I was shown the ad, and was told demographic information, as well as a data partnership with DLX. I would like to know if others who see and ad they suspect was displayed due to microphone captured data have this DLX data supplier source in common.
How many ads have you seen on FB that matched a conversation that happened earlier?
Isn't this the exact definition of confirmation bias?
My backpack zipper broke when I was walking to work. When I arrived at work I told my coworker I needed to buy a new backpack. She and I talked in her office about what type of backpacks her son liked, did they have enough pockets, are the zippers strong. I never used my phone to google anything about backpacks, never looked anything or searched for a new backpack online. That same afternoon- boom backpack ads in my facebook feed. Totally creepy.
You'd have not noticed if you saw a backpack advert at some other time. Just how many adverts do you see that aren't standing out? Huge amounts, probably.
This could be quite easily explained by the frequency illusion.
I'm sure most of the time their algorithms make guesses like that it doesn't result in anything relevant, but you don't notice those ones.
If this cross-referencing of other people's searches is something that they're doing, though, that's concerning too. I don't want my partner to see ads for a gift I'm surprising her with because a facebook algorithm is guessing she might be interested in ads about things I've googled.
https://developer.qualcomm.com/blog/using-qualcomm-hexagon-s...
"Industry wide adoption"
2013
The rest is PR.
> enables devices powered by Snapdragon 800 processors to be “woken up” by a custom voice command (beginning with a custom phrase set by the device OEM like “Hey Snapdragon”, for example)
Because they are always on and have to use minimal power, these chips are used to recognise a very specific term (probably why these companies don't let you customise how you summon the assistant), and cannot do anything like voice transcription, network calls, ... (which would just kill your battery).
Early 2014, phones that used it were:
http://www.freewaresquad.com/2014/02/18/phones-with-snapdrag...
Today that CPU series is everywhere.
The fact that there is no hard evidence makes me very skeptical.
Every anecdote is explainable another way. Someone in the comments said they were having a conversation about storage units and saw an ad for storage units immediately after. Even if they had never searched for storage units -- and can we really be sure? That would be easy to forget -- Facebook might just have extrapolated they they would need a storage unit from related searches, like searching for a moving company.
I would be willing to bet this type of algorithmic extrapolation explains every one of the microphone-ad-conspiracy stories.
I'll throw another layer on the conspiracy pile and say they deactivate that behavior for anyone they notice trying to prove it, similar to Uber's Greyball.
which ones do you mean?
It's surprising to me that people here are still debating this. I believe it's been going on since the original iPhone came out after a friend told me a very specific product appeared for her on Facebook immediately after she spoke about it and went to go look it up on her phone.
I don't think Facebook would stop at any measure to get all the data they can from you if they can get away with it and frankly anyone who does is extremely naive. It's right there available to them. There has never been and may never again be a time when companies (particularly Facebook and Google) to obtain as much sensitive data as is available to them from us without our knowledge, without external oversight, and without repercussions due to the lack of legislation.
It would be great to really inspect the traffic and see if this is really the case. Independent of what Facebook claims.
Either way, I am happy I have deleted it off my phone. I have never been happier each day not reading garbage and actually focusing my time on apps that improve my life and move my life forward.
For example, why use the word 'recording' when the problem is accessing the microphone at all... Facebook doesn't have to 'record' to stream the audio to their server for audio recognition without 'recording' anything.
Of course all of this is just from the title of the article but the paranoid will pretty much always be able to find wiggle room.
Maybe I'm just having a bad day.
And the general sentiment appears to be very far from what you claim. I think it's possible that this issue is going to come to a head this time around. I don't know what the outcome will be. Just saying it's confirmation bias (the likely real explanation?) may not be enough to satisfy people.
https://gist.githubusercontent.com/quellish/473f513fbd131023...
All that said, I still lean towards confirmation bias.
Given that, it's fairly probably that you will be shown ads about things that you are interested in, and it's fairly probably that you'll talk about these topics with friends. Given this and the number of people that use Facebook, it's probably not that unlikely that your experience happens simply by chance fairly frequently.
What I think they do is, they record everything using the microphone, and they store it if you are a person of interest. A person of interest could be that maybe you have a over 1000 friends or something, maybe your a ceo of some tech company, maybe you're a lead programmer at intel, maybe you're someone the US government wants to track, maybe your a senators staff member, maybe you work for the Skunk Works at Lockheed, maybe you said a key word/phrase that throws you into the 'person of interest' category like 'electrogravitic propulsion'. When these surveillance companies do this type of stuff, they do it to people of interest, and if they ever need to put the squeeze on someone, or steal important information like get secret information about a stock trade or something, they go to that person of interests microphone records (which probably get transcribed into text by some program) and search for whatever they want to know. This is why this shit should be utterly banned, it controls people, it steals from people, it destroys innovation and scientific progress before they ever get off the ground, and its just plain manipulative.
If you were just talking to a sample including a small fraction of your friends and acquaintances, and you found that like 10 of them had noticed that Facebook seemed to be showing ads based on their conversations, that would be compelling evidence that something was up. And since for the vast majority of human history, that was the kind of information you had access to, that's the way we tend to process things.
But now you can go online and mention this weird thing that happened to you, and a thousand other people will all see it. Suddenly you have super high odds of a "cold hit", even though the phenomenon was purely random.
And to make matters worse, it's not purely random. Facebook has algorithms that are trying, often successfully, to model your life based on the information you give it. So the chances of "eery" ad targeting are significantly higher than if it were just random.
I think we're going to see this kind of conspiracy theory more and more as time goes on, and we need to figure out cultural tools to counteract this problem.
And not to mention how many engineers that work on Facebook that could be whistle blowers.
Seems like the internets has taken these serious accusations as some sort of corporate conspiracy but I haven't seen any independent, credible study of how Facebook's apps (confirmed) access to microphone are actually being translated into ads.
If this is happening then it must require some amount of CPU and memory use on the device, which even if hidden from the OS can be detected via power use and heat dissipation. It should be possible to indirectly measure a phone's CPU load (including any hidden processors) by filming it in the infrared and/or by attaching it to power and measuring the wattage consumed from the supply.
If this is happening it must also require some network I/O. Something must be phoning home and reporting the data if the data is to be useful.
Note that there is an intrinsic trade-off between the heat dissipation angle and the bandwidth angle. If the device is doing parsing, speech recognition, AI, etc. locally then it must be consuming more energy and dissipating more heat. If the device does these things remotely it must upload more data. Compressed audio is far more bandwidth intensive than digested meta-data.
This seems like something security researchers can and should look into.
People started discovering that after iOS introduced the battery/power usage section in the Settings app.
That's why I'm skeptical that Facebook would do this -- it seems like a terrible business decision given the high chance of getting caught, either by reverse engineering or a whistleblower.
These kinds of cosmic coincidences happen to me and everyone else every few weeks, and they mean absolutely nothing. We're just wired to see patterns. I (and millions more) see hundreds of Reddit posts per day, you (and millions more) see hundreds of ads per day. A crazy coincidence every 10,000 ads isn't very strange.
Something like this should be piss easy to prove, especially on Android. Traffic monitoring, monitoring when exactly the microphone is accessed on a rooted phone, anything really. The fact that this rumor has been going about for years and nobody has found anything, means it's just nonsense.
One could argue, from a PR perspective, it does not matter whether Facebook is listening or not.
First, some users do not trust them. Maybe Facebook is worried about those users telling other users.
Second, some users believe it is technically feasible to listen.
Finally, there is no legal/regulatory framework in the United States to effectively stop Facebook from listening[FN1], if they wanted to do so. They have clear business incentives to listen because the objective is to gather more personal data about users than their competitors.
NB: Added "in the United States" for clarification.
FN1. Incorrect? What are the theories of liability? What are the sources of law or regulation in the United States? Is injury required? What is the injury? How does one prove the injury?
Edit: added missing word
However, I still turn off microphone access for facebook and instagram. It's probably not fool proof, but it can't hurt.
I doubt Facebook does it with their app. It's too much trouble, blatantly illegal and easy to get caught. It's most likely malwAre selling the info to fb, who has no idea how any advertisers get their info.
If I was FB, I'd either process the audio on the device to avoid network traffic, or buy the results from a partner, to keep the actual listening at arms length.
It wouldn't be that hard to imagine another ad-tech company collecting the data in some other app and feeding the data back to FB. I can't imagine FB would be asking as many questions from a 3rd party as it would to internal teams.
https://www.reddit.com/r/videos/comments/79i4cj/youtube_user...
Why has nobody ever been able to find actual evidence in the code?
A girl I was seeing years ago was looking for a new apartment. She texted me telling me she found a new apartment on 20th and Gladstone Ave and asked me if I wanted to go with her to see it. Minutes later, I get a Google Now update saying "It would take you 43 minutes to get to 2020 Gladstone Ave if you leave now".
I thought it was cool at first - Google Now was getting really helpful with the driving suggestions. I also thought it was strange that it was for a specific address - I thought that was a glitch, or the nearest address to that intersection, although Google Now has worked with intersections in the past.
I thought Google telling me the time it would take to travel to her apartment showing was impressive enough to actually tell this girl what my phone did. She was not amused and thought that I had hacked her phone. She wanted to know how I knew the exact address of the apartment!
I was curious how this could have been possible since she never texted me the exact address and found out she did look it up on Google Maps. So it's clear how Google had the data but sharing it with me seemed like a huge breach of privacy.
Perhaps the craziest part of all this is that not 5 minutes after telling this girl what Google did, when I went back to take a screenshot of the suggestion - the Google Now card CHANGED to 20th and Gladstone Ave!
I was seriously wondering they were analyzing my conversation afterwards expressing my concern of this breach of privacy and had the "privacy invasion" level turned down a degree.
But I also know corporations are usually scrupulous about following laws so unless there is something in Facebook's ToS/EULA that mentions doing this I'm skeptical.
I don't see why the vague terms in their data policy [1] couldn't include audio (all-caps are my own emphases):
> We collect information FROM or about the computers, PHONES, or other devices where you install or access our Services, depending on the permissions you’ve granted. [...] Here are some EXAMPLES of the information we collect:
> - Device locations, including specific geographic locations, such as through GPS, Bluetooth, or WiFi signals.
If device location is merely an example of information they gather from my phone then why couldn't audio just as easily be another example?
Could this become the next Volkwagen?
Facebook denies 'listening' to conversations | https://news.ycombinator.com/item?id=15581114 (Oct 2017 20:30, 171+ comments) http://hnrankings.info/15581114
Basically the same discussion on separate stories covering the same topic, both on the front page simultaneously. The second only took off after the first was flagged off the page or something.
I didn't want to summon the dupe police before I had a chance to review it a bit closer. Also didn't want to lose both to the flagging committee.
Mostly planning to review participation by various HN users. #facebook #haterade
PS. Neither one is worth an up-vote from me, let alone adding as a favorite!