Ask HN: Has anyone tried to debunk the myth that Facebook is listening?
For weeks and months, there has been the same conspiracy about Facebook listening to their 2 Billion users to serve targeted ads.
Countless questionable youtube videos, reports of all sorts by people who have no idea what it would entail to perform such a stunt on a large scale.
After googling extensively, I still can't find any real scientific proof that it's either true or on true. It's mostly about some people relating their funny experience and mostly about Facebook saying "No we don't"
Anyone familiar with Charles Proxy would know it's pretty easy to spy on the traffic going from an app to the backend. I work in Facebook advertisement and I've been doing it for various reasons already many times. I've never seen anything that would like a speech datagram, but then again, it wasn't what I was looking for when listening to the Facebook app.
I would expect that given Google and Apple should be scrupulous about what the Facebook app does and doesn't, there should be a host of people who've been looking into that urban legend a number of time. Also the data plan of many people would simply explode too quickly for this to be true.
With my colleagues, we've been wondering if it would be possible to rely on a local text-to-speech engine running on the phone. Needless to say that would drain the battery pretty quickly and the dataset needed in the app would be pretty huge... But is it? And would it really drain the battery? More than Pokemon or the Facebook app...?
Who among you guys is a seasoned user of Charles and could run the experiment a tad further? The lack of actual responses to your question here is a bit entertaining. "Of course they are not", "it's not possible", "iOS/Android is safe" etc. Almost like it is offensive to ask. I can't help but remember my incredulous reaction to ECHELON in the 90s and how it turned out to be not just true, but much worse than the original "conspiracy theory". A very relevant example: https://jscholarship.library.jhu.edu/handle/1774.2/36569 This post was built to influence the largest number of people possible. The title is worded to reach people that don't click with the suggestion that the accepted view in the community is that they aren't spying. It's also by a newly-created account made by a Facebook Ads employee. The microphone privacy invasion rumor has spread so far without evidence because people are tired of the 1000 other privacy invasions for which there is evidence. They don't care whether or not this specific instance is true. People are tired of FB being openly scummy at every other turn. FB is the least respected tech company for a reason. Happy to debunk this in whichever way you want! I definitely don't work for Facebook. Also I have been consuming HN content for years without contributing. There has to be a first day... I do work closely with Facebook (For a Facebook Marketing Partner called Smartly.io) and we collaborate closely on certain aspects.
I think I have a very deep understanding of what Facebook is capable of and that's the reason why I want to debunk this.
Also, I am really tired of hearing the questionable reports. Look me up if you which: https://www.linkedin.com/in/momeunier/ can't help but agree with you. people are scrambling to debunk an ill-defined phenomenon that's only exists anecdotally... and in the realm of those who ask "qui bono?" who benefits if "facebook is listening"? what are the incentives? the question of technical possibility is the first place people will go to... without realizing that if there is a will, there is a way. Few question the feasibility of building a mechanism. The question is whether they’d try to hide any such mechanism in plain sight. Native apps on a phone are not comparable to government spying on networks outside of your control. The attack vector in this case is entirely within the device in possession: the code, and interactions with the OS and hardware. Entertaining the opposing view: If fb did build this, how is there no hard evidence of it? On iOS, their code is distributed via Apple, so they couldn’t doctor versions for specific users, and any attempt to build self modifying or self deleting mechanisms would be detected sooner or later for similar reasons. That’s not what’s happening. There’s no hard evidence and it’s not really a fair comparison between detecting government snooping on networks out of your control, versus a spyware mechanism hidden in plain sight, entirely in your pocket. Nobody seems to be doubting the feasibility of building such spyware but rather are asking where’s the hard evidence and how could they get away with it (not legally necessarily but in terms of detection)? My answer to this has always been "Why would they need to?" It drains battery, it sends a lot of data, and would be a huge scandal. But you are already telling them where you are, who you're with, and they have information about your credit, etc. Instead, I've been using this as an opportunity to teach people why big data is scary. In the past people often said they didn't care that Google could, and do, read their emails. Often responding with the ironic "I have nothing to hide" quote, or "I don't care, my stuff is boring." This pervasive rumor, of Facebook/Google always listening, has facilitated this conversation immensely. I've been able to better explain it to many who are not tech or statistic savvy in the least. I don't believe Facebook is listening, because they don't need to. Or in other words, they are listening, just not to your voice. Their product is an intelligence-gathering apparatus that works at scale without microphones or other traditional surveillance tools. I'd phrase that as "they're watching" not listening. More like having a private eye tail you. They can't hear exactly what you say (well maybe) but they sure do see everything you do. I try to put my mind in the head of whoever at Facebook would be making this decision. Chief of marketing? Chief of Technology? Mark Zuckerberg himself? A product manager? Some director? And I believe all of them in fact would be perfectly fine with it - if it somehow helped them hit their quarterly bonuses or user growth numbers. My point is that I do not believe constantly listening would significantly improve their profit. I don't think they'd gain much information. And I'm sure they've discussed it. > it sends a lot of data They could send just a transcript once every few minutes. You probably wouldn't do that sort of processing constantly on device. Battery intensive is it possible that there are tools available that do the job with minimum energy consumption but they are not out in the open yet? i wonder what tech military has that we dont know is even possible today? The military is not at the forefront of R&D, but they may be able to restrict whatever they deem important. For instance, this can happen with patents. Some patents are reviewed by a committee before they are published. What do you do about multiple languages? Hiding a speech-2-text dataset in English in the app is one thing, hiding several in multiple languages would make the app bloat. I was at lunch with a couple friends and they were making fun of Juicero. I hadn't heard of it at this point. I didn't look it up on my phone, but at least one of my friends did. That afternoon I started seeing Juicero ads non-stop on Facebook. Was Facebook listening or did they connect the dots that my phone was in the same restaurant as friends who were Googling Juicero and figured we had been talking about it? Another time I was on a road trip with a friend, driving through a food desert. We were talking about our food options -- stuff we never eat like McDonalds, Wendys, etc. Guess what pops up on Facebook moments later? An ad for Wendys! Were they listening? Or did they just connect the dots that it was lunch time, that we'd been driving for a long time without eating & that Wendys was one of the only options around? In both instances, there were non-microphone explanations but creepy/impressive nonetheless. There's also the it's all about how you react explanation: brands blanket people with advertising which you wilfully ignore except for those rare instances where it coincides with something you want or something that happens to resonate because you were talking about it moments earlier. And maybe your friends were making fun of Juicero at the time after being prompted by being bombarded with ads for it, rather than you being bombarded with ads because your friends made fun of it... I think I've got a lot of Facebook ads blocked, so I'm looking at a right hand column which is recommending me pages of random companies one of my friends has once liked and suggesting translations into two languages I don't speak which are spoken in countries I've never visited and have minimal connection with. Which I guess is an upgrade on all the dating ads which didn't even bother to match my "interested in"... >Guess what pops up on Facebook moments later? An ad for Wendys! Were they listening? If you saw an ad for Wendy's in any other context, would you have been surprised? I saw an ad for Wendy's shortly before coming to comment on this post. Are HN's advertising algorithms predicting the future, or is it just a coincidence due to the overwhelming volume of fast food advertising? I can't remember seeing one before & haven't seen one since. "Feynman's Grandmother"! I bet Wendy's also had a billboard on the highway leading to that food desert uh? They know people are on their phones looking for food options (since clearly they aren't easy to find) so they target that area with their ads. It's not that hard. This is getting similar to horoscopes. When they are wrong, whatever. When they are right, it's "creepy". Every one of my friends, including me, has a story about how they used a phrase in conversation that they don't typically say ("Renaissance Faire" in my case, which I have never attended and generally have no interest in), and then seeing an ad for it on either Facebook or Instagram right afterwards. This is purely anecdotal, of course, but it is creepy anyway. The only way to stop this "myth", if it is one, is for Facebook to stop doing whatever it is they do instead of listening to us to get to this unnerving level of targeting. Alternate explanation: the person(s) you discussed this with did some google'ing and browsing. The tracking cookies do their thing and since you and those persons are linked, you end up with the ads served. a good alternative hypothesis, but also one that should be proveable, at least in some cases. designing an experiment to test this would be quite easy. ultimately your explanation could exist alongside audio snooping, perhaps even helping to resolve areas of ambiguity... Is this repeatable though? It seems like this could be explained by a combination of the Frequency Illusion / Baader-Meinhof Phenomenon[0] and the Birthday Problem[1]. That is - it is possible that the reason you were discussing the topic is the same reason that advertisers were targeting that particular phrase at that time. The Frequency Illusion could explain why you noticed that particular ad and why you were surprised. Also you presumably discuss topics all the time that you don't see ads for, but the probability of seeing an ad for any of those topics is quite high (the birthday problem). [0] https://english.stackexchange.com/questions/153166/what-is-t... 100% - I'm not sure why others are so dismissive. It's not that technically difficult, and FB has a reputation for being creepy. If only I could figure out how to get random one night stands out of my "People you may know" list... The excellent "Reply All" podcast had an episode about it: https://gimletmedia.com/episode/109-facebook-spying/ It’s trivial to debunk by observing that many people who say they’ve experienced this are using iPhones. On iOS, apps can’t silently use the microphone in the background. A background app using the microphone is always accompanied by a red status bar saying that they’re doing it. This doesn’t show Facebook listening, therefore we can conclude that they’re not, at least on iOS. It’s possible they might be doing it on Android, but once you show that a lot of people are imagining things, it’s a small leap to thinking that they all are. > It’s trivial to debunk by observing that many people who say they’ve experienced this are using iPhones. On iOS, apps can’t silently use the microphone in the background. The same way people believed iOS background apps couldn't record your screen? [0] When you cannot trust the operating system manufacturer to be impartial, it stops being trivial to debunk. [0] https://gizmodo.com/researchers-uber-s-ios-app-had-secret-pe... Android 7.0+ does ask for permission to use the microphone when requested, and can be turned off in the settings. I believe this is also on 6.0 but I'm not 100% sure. Wasn't there just a story about Apple giving Uber special permissions (which were not turned off after they were no longer needed) so they could do stuff on the Apple watch? Is it that much of a stretch to think they would also give Facebook special permissions? Yes, it is that much of a stretch. Apple temporarily gave Uber that special permission in order to make their Watch app work better for their users. It’s extremely unlikely that Apple would give Facebook special permission in order to spy on users to no benefit to the people who actually pay Apple money. And somehow it’s gone unnoticed all this time. I didn't realize it when I wrote the above, but this makes my case much stronger: these special entitlements are listed as part of the app. It's far from obvious to the end user, but people who know how to look at them can do so. Either nobody knowledgeable has ever looked at Facebook's entitlements, or it doesn't have one that would let them do this. given the below examples i'll contest that it'd be "trivial" to debunk on the iOS platform and also add that even if it were debunked on the iOS platform, that doesn't debunk the other platforms. why would they leave data on the table if they could scoop it up instead? All I know is that things that I have talked about with my phone around 1-2 days prior, I suddenly see ads for. Things I did not Google or mention on the Internet. In one case I started seeing ads on Instagram for a product that I posted a photo of and did not even include the product's name in the post. Dishwashing liquid that I would never even buy (I'm not the target user so it wasn't target marketing). It's all a little too much for me to believe it's a coincidence. The burden of proof falls on the people making claims that facebook is listening. Absent that, it seems extremely unlikely that facebook is recording audio and no one has posted in outrage about it - streaming audio would be quite conspicuous. Even more troubling than audio recording in this case is that facebook has so much information about you that they don’t need it. The idea that facebook knows the approximate content of your conversations/thoughts to the point that it seems like they’re listening should be deeply troubling. If facebook knows the content of your conversation, seemingly before or at the same time that you have it, what does that mean for free will? Agency? (Obviously free will is bs, but most people I’ve talked to refuse to acknowledge that - is this a wake-up call?) What other predictions can they make? With the level of information given to facebook (browsing history, precise location, when you’re asleep and when you’re awake, and possibly much more, like your cell account info and the TV you watch) they can build a very complete picture of your life, and may even know things about yourself that you don’t know. Of course, when facebook is actively listening to you (e.g. for facebook messenger) all bets are off. They already read your conversations anyway, so it’s not a far jump. I don’t think FB is listening because all the stories I’ve heard can be explained by their graph. I think it’s hard to debunk because people don’t recognize how powerful their knowledge graph is at this point, and simultaneously underestimate how hard general-purpose speech recognition is, even with studio-level audio setups. And especially to a non-programmer, I think it seems so much easier to just figure out speech recognition than to maintain a huge knowledge graph about billions of people, their connections (explicit and implicit), their current whereabouts, the sites they are visiting, and the text they literally give to Facebook in the form of messages etc. But of course, each one of those things is a tractable problem, and can progress independently of the others, until all of a sudden it’s so good it seems like they must be inside our heads, or at least listening. But here’s the thing: I think all that big data is actually waaaaay creepier. If they were just listening in and showing relevant ads, that’d be one thing. But instead they have all this data going back years, and can make all these inferences from the data. Just imagine what inferences could be made about non-advertising topics! I proposed a thought to Reddit, but have no proof of its possibility either technically or practically. The people we talk with are generally the people we are connected with on social media. It's feasible to me that a conversation happens and someone in the group makes a search, which then maybe causes algorithms using location data to push those ads to people in that location. Further, what's being talked about is probably being talked about and searched by others in the area and/or with the same interests, so even if it isn't localized quite so much the possibility of what's being talked about to generate ads is heightened. Finally, there's the confirmation bias and recency bias at play. Once something is known or on mind, it's more likely to be noticed. They may have already had those ads coming in and only noticed them after talking about it. Or, the ads might have even caused the person to subconsiouciously think about the thing. The psychology at play in advertising is advanced. "Facebook does not use your phone’s microphone to inform ads or to change what you see in News Feed. Some recent articles have suggested that we must be listening to people’s conversations in order to show them relevant ads. This is not true. We show ads based on people’s interests and other profile information – not what you’re talking out loud about,” said Facebook in a statement last year. “We only access your microphone if you have given our app permission and if you are actively using a specific feature that requires audio. This might include recording a video or using an optional feature we introduced two years ago to include music or other audio in your status updates.” Source: https://www.forbes.com/sites/amitchowdhry/2017/10/31/faceboo... Facebook listening to their customer's concerns about privacy? Yes, that myth has been throughly debunked. Maybe local text-to-speech is not plausible given the heavy processing and battery drain. What if instead there is a list of words/brands being listened for; my phone is always listening for "OK, Google." I can't speak for Android, but there are a number of precautions on iOS. The first one is that iOS gives users very fine grained control over what apps are allowed to use the microphone. It's not just a list of permissions, it asks specifically with a system dialogue box whether you want to allow an app to have permission. Second, the status bar turns a very obvious red color when an app is using the microphone when it's in the background. If FB was able to bypass both of those safeguards without Apple knowing, it would be a major story. Install the facebook app on lineageOS, open up privacy guardian, and see for yourself exactly how often it uses your microphone. It's definitely not listening to everything you say. Just pointing out that this is far from a conclusive invalidation technique. Too many ways such spyware could still exist or operate. i realize i'm fighting a losing battle by posting in a conspiracy theory thread, but okay: what are those ways that the facebook app could listen to you without the OS knowing it's using the microphone? I don’t know, but one would need to thoroughly audit the code to find out one way or another. One would not draw conclusions about the non-existence of such a feature just because they’ve made exterior measurements (exterior to the black box blob of fb code) that don’t give affirmative indication. Contrarily if the technique you described did uncover such surreptitious microphone activity, that might be a bit more conclusive. (Hard to disprove until you’ve considered all ways it could occur, and reviewed each one, and even that leaves possibilities for mistakes in method or theory) The viewpoint is that there’s no hard evidence such behavior occurs from their code, not that it couldn’t be the case (still). Maybe some revelation will come out but I’m not betting on it. Fb is too smart for that and their official explanation makes sense. For those that “believe”, this lesson never gets old: https://en.m.wikipedia.org/wiki/The_Thing_(listening_device)
In this example, if your theory for locating a bug was based on the requirement of an electrical signal, you’d have missed such a bug entirely. Next to impossible therefore to disprove such a bug exists, if you can’t consider all of the ways in which it can exist. Hope that helps you on your mission to learn reverse engineering (a skill sadly missed in the USA throw it away culture) Without hard evidence (speech processing code, or evidence of the recordings over the wire), it’s reasonable to take the official account at face value. Frankly that account is far creepier. They’ve built AI so smart that it knows what you’re talking about. Think about it for a minute. Fb and others have so much data on people and have trained algos so well, even they don’t understand how they work so well. It’s not difficult to imagine this is the case; a computer, especially one with extensive data on the lifes, behaviors and thoughts of billions of people, is far smarter and better at predicting behavior than any human could. Scary is that it’s just the beginning. I fall squarely into the alarmist camp on the AI issue. Not now, but in the long run. One can see how this scenario with the fb ad microphone spying paranoia thats actually 1 million IQ AI flexing its muscle is but the beginning of a slippery slope. Facebook may only need to recognize a few keywords. The one that companies had payed for. It wouldn't be that many. An approximate recognition is good enough. So processing and data transmission could be small. As long as there is no uncontestable proof, the doubt must benefit Facebook. The testimonies are not uncontestbale proofs. Considering the million people uisng the facebook app, there will always be a small set of people who will see advertisement of what they just talked about. That is statistic. Some of these, understandably chocked, may be very vocal about it. We also know that people are payed to spread fake news and manipulate people. So I'm cautious. The truth is unknown, and the longer it takes to get uncontestable proofs, the less credible the listening thesis becomes. Facebook can't do much appart to let it rain. Open source is the way I would go. That is one way to calm down paranoid or conspirationist people. >As long as there is no uncontestable proof, the doubt must benefit Facebook. they have a history of privacy violations they have an incentive to snoop further i see no reason to offer the benefit of the doubt to either hypothesis, as anecdotes do not constitute data but for what it's worth, my hunch is that there's some extra snooping going on that has yet to be unearthed. might not be audio, necessarily. Interesting, curious to hear more. What techniques are you referring to? Some re-appropriation of mobile hardware for cross sensory applications? Data sharing of some kind? I believe that facebook in no way can accomplish this at scale without getting caught. Even so, i recon that your approach to test them is flawed, instead you should either of the following: 1. Reverse engineer the applications that has access to actually capture audio from the microphone to start with, i.e. the Messenger app on android. 2. Create a kernel module/microphone driver, which records when and which applications actually access the microphone. If the facebook apps are actually found accessing the microphone outside of a user initiated scope, one could bring the experiment to the next level and record the same audio recoding and save it for later review. I recon 1. has already been done by several researchers, especially in relation to bug bounty programs etc. but that doesnt really mean that you should try to do it again for your own research. I think this is really something even more nefarious than background listening. Facebook and Google are using machine learning and massive data sets about the mental worlds we inhabit when we think nobody is looking. And these algorithms are optimizing for clicks in so many vectors of psychological need and vulnerability that it’s already become incomprehensible for how they can know the things we haven’t yet understood we will think about. They know about our conversations because they know the searches of the people we talk to, and what they’re thinking about. What will be on the top of their mind. There’s nothing to go looking for or debunk. It’s just the next step in the data collection experiment. I’m guessing they’ll need to tone down the effectiveness of ad targeting, if they haven’t already. for those of us who are digital natives, even if we spend great effort in covering our tracks, they likely know more about us than we know about ourselves. they use this information to employ a system of psychological levers which results in us willingly handing their proxies our money. is it a good business practice. certainly. is it simple theft? my instinct is to say no, and instead ask: is this new mastery of knowledge about peoples behaviors and the ability to affect other behaviors on a mass scale the most terrifying form of feudalism that mankind has known? Reply All, an entertaining podcast from Gimlet Media, tackled this topic a few weeks ago [0]. They explained the myth, explained how Facebook's advertising an be explained without surreptitious listening. In the last part they called believers in the myth and tried to convince them otherwise. Anyone with a heavy accent have these 'coincidences'? I imagine that a thick accent (Scots,etc) would be much harder to have this happen to. I personally find the alternative far creepier: that Facebook can figure you out and predict your behaviors even if it doesn’t actively listen. There certainly is a double standard. I.e.: it is creepy when FB listens to you conversations. But if they use AI to derive their knowledge about you that somehow makes it alright. For me personally, the fact that they can learn so much about me is what's creepy and reason enough to not use their services. The how is merely an interesting bit of trivia. is there any evidence cameras are used? I've recently been engaged in two activities in front of my phone and got scarily relevant ads that I've never seen before, and are can not be part of my graph. I am not on social networks and none of my browsing or search would be vaguely relevant. It’s really amazing that people think Facebook is listening. It’s like wearing a name tag and assuming someone is psychic when they call your name. This is pretty much impossible to debunk because: A) Given Facebooks policy of A/B testing, it almost certainly isn't listening all the time. and more importantly, B) Now that everyone is talking about it, Facebook would have been stupid NOT to turn it off at least for now. Would they even take a chance on this with anyone knowing it’ll eventually be discovered or reverse engineered? Why not take their account at face value? (We’ve built AI that is so smart that it can predict your conversations). >(We’ve built AI that is so smart that it can predict your conversations) We've built AI that can do a great job of slicing and restitching your old conversation into responses that are plausible responses to input sentences. Whether that's good at predicting a conversations depends a lot on the conversationalist you're attempting to emulate using that technique. Volkswagen Thought experiment for both scenarios: I’d guess that my emissions cheat code blob is “safe” since no one is looking. (That guess was wrong of course). I’d guess that my fb app code, installed on billions of phones and scrutinized by researchers the world over isn’t going to be able to hide anything related to processing or transmitting speech recording surreptitiously. Thought experiment for both scenarios: Every significant national government has a major agency in charge of protecting the environment from harmful emissions with many paid engineers devoted to emissions testing. None of the major national governments except Germany have an agency devoted to ensuring that devices do not en fringe on user privacy. If you think that Volkswagen is a bad example take AstraZeneca [1]. No one had to prove that they had committed fraud by missmarketing their drugs. Much of the marketing material was publicly available. Everyone who knew the law, knew that they were breaking it. Or perhaps you could look at Seimens, which outright bribed, using bags of cash, government officials, in order to close deals [2]. Moral of the story, corporations break laws. If Facebook is listening to you using your phones microphone, that's not even illegal outside of Germany. [1] http://abcnews.go.com/Politics/Health/astrazeneca-pay-520-mi... [2] http://www.nytimes.com/2008/12/16/business/worldbusiness/16s... Valid points but doesn’t address the basic idea that fb knows their code is widely scrutinized. If they process or transmit audio surreptitiously, or have done so, they’d be found out. Their app is so pervasive that it’s exceedingly unlikely that they’d be able to hide any such mechanism for long. In regards to B) there would still be a paper trail, in terms of old app versions etc. for ever available. The people who believe it’s listening would never accept even the most ironclad “proof” that it’s not. This is a hallmark of conspiracy theorists. So why bother? I believe it's listening based on an overwhelming amount of anecdotal evidence reported all over the Internet. (Where there's smoke, there's fire.) Including by intelligent, scientifically-minded people. At the same time, it wouldn't take much at all for me to believe otherwise: a simple blinded test where some people leave their phones in a room, researchers come and talk about some subject (absent the owners), then leave without ever interacting with the owners of the phone. There are two ways to do the next part: The owners of the phone 1 week later could be asked if they've seen any facebook ads on (topic). A control group would have to be compared (where researchers didn't speak on the topic.) Alternatively, the subjects of the study could be given 5 random topics when commencing the study and asked to record any ads they've seen on any of the 5 topics. The researchers could then speak on 1 of the 5 topics (differing from person to person) and a statistical analysis could be performed. A "proof" is a statistically valid correlation between what the researchers talked about out and what the blinded subjects reported seeing ads about. A refutation is a lack of such a correlation. Easy, and it would convince me personally. By the way based on the anecdotal evidence I strongly expect this study to conclude "facebook is listening." - Note: there are a few loose ends to take care of. The researchers who are tasked with speaking on a subject are more likely to google it on their own phones. (Having been exposed to the topic), and those are in geographic vicinity. There are other similar possible mechanisms. Perhaps the best approach is if subjects' phones are in a sound-proof vault and the researchers' speech is either fed into it via speaker/microphone, or not done so, but the researchers do not know for any specific phone whether it is able to hear them. (Making the study "double-blind", as neither the subjects nor the researchers know whether the subjects' phones have heard anything on a subject.). I don't agree with your proposed way of studying this. You're trying to study a symptom and then conclude on the cause of the symptom. That makes no sense. If you want to prove that Facebook is listening, then prove that Facebook is listening. And don't try to prove that Facebook is listening and then targeting you with ads based on what you said. There are way too many levels of indirections that can trigger false positive for a vast number of reasons you don't control at all.
What you are trying to do is replicate in large-scale displays of anecdotes but with a slightly more controlled environment.
That won't prove anything since it will just be anecdotes and will again not sustain rational explanation by experts of Facebook ads mechanisms. Without getting into the ads delivery part and the anecdotes, how would you prove that Facebook is listening?
How would you prove that there is a set of information taken from your speech or your audible environment transferred to Facebook. I think you didn't read my proposed methodology carefully. We can pick subjects such as: It is important in order to maintain double-blind environment that the researchers not hear whether they are being amplified into the box. The results might potentially look like this: Of course, I just made this up. (I imagine the subjective 1-5 scores being whether the given subject reports seeing such an advertisement, from 0 definitely not to 5 definitely yes.) I even made subject 3 unsure about topics 1 and 3 to mimic that humans are fallible. Likewise subject 2 does not really report any advertisements. (This is likely in the real world - for example subject 2 could be explicitly excluded by advertisers for some reason.) The attached is the kind of graphs that I would expect based on dozens of scientifically-minded people trying them. If these are the two graphs that we got, and if the test and control groups were truly randomized, what other explanation could you offer? Of course, my proposed experiment is orders of magnitude more scientific than what people are doing with their n=1, unblinded personal experiments. But theirs has some validity also. Thanks for explaining further. That would actually be an interesting experiment. Who's going to run it? Not sure who would run it. Nobody really cares that much. I disagree with your feelings about the anecdotes given there is an official account that’s more likely. Why would fb come out officially saying no, knowing they’ll be uncovered eventually if the answer ever were yes? Also, your techniques could reveal a false negative both in the existence of a bug in the particular build of fb code being tested, and that such a mechanism could have existed in previous versions but already removed. Likelihood has to be judged against the evidence. A lot of scientifically-minded people have tested it for themselves and found ads to appear when they spoke about subject explicitly to test whether ads would appear. I recall several specific cases. It's not quite scientific enough for me, but I'd expect a scientific study to have the same conclusion. Just like if you do a scientific study of whether baked beans make you toot, yes, you will find that they do. Although you state there is a high likelihood of a false negative, it would be a useful study for me and others in the more likely case that it matches the informal n=1 studies numerous technical people performed by themselves. Source? >Source? [1] https://www.google.com/search?q=%22baked+beans+make+me+fart%... [2] https://www.reddit.com/r/videos/comments/79i4cj/youtube_user... [3] https://www.youtube.com/watch?v=jjbz1N4qBQA the host themselves did the same. In the original comment of mine that you replied to, I suggested how to make the methodology even more scientific with a double-blind study. As far as anecodtal evidence this is going to be the best that you can do. I personally find it convincing. These are reasonable people testing things as scientifically as they can. Dude, cmon. I know this is hn and I’m supposed to put effort into crafting useful responses. But, how have those links convinced you that it must be the mic? From the evidence in those videos, I’d rather be considering whether fb is part of a future AI continuum monitoring and influencing the past (from the future in which the AI exists) by messing with people testing this feature. Yep, cat is alive on Facebook/IG, days later. Microphones and big data/AI are but two slits through which photons can pass, in the experiment known as fb. What other kind of explanation do you expect? People have tested it by playing a Spanish soap opera on their TV for a few hours, leaving their phone in front of it to listen. They then started getting Spanish advertisements. I realize that there is a confirmation bias in that they are then actively looking for Spanish-language ads (to test their theory) but don't you think they would have noticed them at other times? They explicitly ran an experiment and got the verification. In fact, some people were incredulous and re-ran the experiment thinking there's no way they would start getting Spanish ads: they did too. That's all it takes to convince me. I mean what other mechanism would get back to Facebook? As far as I know TV's don't report back which specific channel they're tuning into (all channels are streamed concurrently.) I'm all for a more scientific test but at the moment I find the anecdotal evidence quite convincing. I just outlined how a scientific test might work here: One has to be very thorough when reviewing the details of such investigations. For instance, we would want to look at details regarding the experimenter’s cable tv service. If it’s digital perhaps the viewing data was resold to fb. At the most in those cases there’s enough info to suggest data sharing is going on at the least, but not enough to confirm anything regarding the fb app itself and the mic. Worth investigating more, though. > it wouldn't take much at all for me to believe otherwise ...goes on to describe multi-week, double-blind trials involving human subjects. :D I get your point. The specific reason I said it "wouldn't take much at all" to convince me otherwise is that under a negative result, we do not know: -> Was Facebook listening under previous versions? -> Did the researchers adequately trigger ads? -> Are subjects able to notice all direct ways in which listening data is being used? Instead, I would simply accept a negative as showing that no, Facebook isn't listening. (I'd happily accept false negatives.) Instead, what is really interesting is that numerous people decided to do an informal experiment such as put their phone next to their TV set to Spanish-language programming and then, to their shock and horror, Facebook began showing them Spanish-language ads. >an overwhelming amount of anecdotal evidence reported all over the Internet This is exactly why I believe pizzagate, am a 9/11 truther, etc. Public interpretation of shared information is not at all the same as anecdotal evidence! Don't conflate these two things. I know beans make people fart because of anecdotal evidence I've heard from different people. That doesn't make this "fact" wrong. On the other hand, interpretation of public information is quite different. - Aside: Also I have a small aside here for you personally. I'd like to remove something from your mind so that you never speak about it again. I don't want 9/11 truthers derided here on Hacker News because I consider the example to be one that should not be included here. Flat earthers? Sure. As for 9/11: I'm not a "9/11" truther. I've never called for an investigation into it. I don't care about it. But for you, if you care about this subject for some reason, the standard for whether you should be a 9/11 truther, in the sense that you believe anyone had any prior knowledge those acts would be committed, is whether you find any leaks to that effect by such a government official.[1] The standard regarding whether it is feasible for 9/11 to have included a controlled demolition is the opinion of civil engineers, whose profession this sort of thing is.[2] So while I can't speak for you, I do have my own standards. This does not mean I'm a 9/11 truther. I'd like you to stop using it as an example, though. Use things that are false (like flat earth) if you want to make some kind of cheap point, please. Note: I want to reiterate that I am in no way, shape, or form, a 9/11 truther. [1] https://www.youtube.com/watch?v=i6eMq5Rit1w - "From about August 23rd, every night, until September 2nd, September 3rd." This insider says "We are never going to look for who they are, and it's probably not worth looking for them." Her Wikipedia page is quite credible as an insider. In addition it includes major smears on her character and a complete refutation over the official channels you would expect. She meets the standard outlined above. [2] https://en.wikipedia.org/wiki/Architects_%26_Engineers_for_9... - "The organization has compiled a list of criteria for a controlled demolition that it says the collapse of the World Trade Center meets: the destruction followed the path of greatest resistance, the debris was symmetrically distributed, the rapid onset of the destruction, explosions and flashes reported by witnesses, steel elements were expelled from the building at high speed, the pulverization of the concrete, expanding pyroclastic clouds, lack of pancaked stories in the debris, isolated explosions 20 to 40 stories below the wave of destruction, molten steel and thermite traces found in the debris." Not all people with such a belief are immutable in that view, nor are all conspiracy theorists. There’s value in debunking a myth that seems plausible to the masses. In this case, the myth is plausible, and enough people without sufficient critical thinking skills believe it, therefore it’s worth debunking consistently at least within the tech community.
The researchers speaks about some subjects, but not others. The ~ represents that in some cases the researcher's conversations are being fed into the box, and in others they are not. The box is otherwise soundproof, and inside is the phone being tested. Box Researcher | Owner of phone (outside room)
[ ]~ x | o
The test group is that the researchers' voices are being fed into the box. The control group is that researchers voices are NOT being fed into the box. - #1 Adult incontinence
- #2 Cat food
- #3 Last-minute trip
- #4 ..
- ..
- #10