I have a Facebook account With literally zero friends (created only to create business pages). On 08/05 I was randomly kidnapped at gunpoint when out of town and wouldn't you know Facebook recommended friending my kidnapper (and some of his family members). Because he took my iPad and connected it to wifi he was located and arrested only 1 day later, before the friend suggestions, but I'd love to know if sans iPad location I would have gotten the recommendation from FB and been able to identify him that way.
Separately I'm getting all kinds of people from that town, I can only imagine why, but it's pretty damn scary. And of course I have to wonder who I'm being recommended to and/or my family is being recommended to.
It’s possible the kidnapper looked your name up on Facebook after stealing your iPad. Just searching for someone’s name is enough for a recommendation.
This has happened with some of my coworkers who I have no mutual friends with and live in a completely different city (didn’t connect to work internet or have GPS permissions on FB either).
Though very uncanny, this is rather easy to extrapolate even using the basic facets FB has available. But the most overlooked is photos. They are offline beacons and are to offline tracking what websites are to online tracking.
For example you and a group of 10 other people took photos at the same location which FB sees as a small gathering of intimate friends. It will need to qualify the location is not a public restaurant to be sure its an intimate gathering. The chances of you connecting with that person are then really high. And if you are and the other person are in each others photos, then it's almost a certainty you will end up connecting on FB should FB recommend the connection. The more connected FB's network is, the more it can extrapolate based on commonalities from the first degree graph of your network.
It's also the reason why Google is betting so heavy on photos and offering free unlimited storage -- remember, there's no such thing as a free lunch. Google wants to build a social network graph based on location and facial recognition to draw on proximity.
With AI this will get even more uncanny. For example a wedding will have a certain photography profile (number of pics taken, the time of day, the location and venue based on Google Maps or past photographing histroy, the lighting in the photos, etc). Once you throw AI into the mix you realize Google doesn't need to draw out any conclusions. It can throw all these parameters into the AI engine and draw up proximity.
In this case, Google or FB will not be able to tell you how they drew the connection, because even they won't know. All you can assume is the AI engine will take dozens of parameters today and hundreds tomorrow. Google's deep investment in AI infrastructure is a bigger testament to this.
As per the author's aritcle it is definitely not a direct connection. FB relies heavily on first-degree connection analysis. If you and I end up in the same phonebook (which the author mentions) or the same photograph and FB connects us, we'll never know why. The not knowing why part makes it uncanny.
Facebook is to people what Google is to websites. It is in the business of enriching its social graph in every way possible and then running profiling algorithms on that graph. Google by contrast crawls and pageranks the crawled graph.
FB has literally hundreds of data sources to enrich their social graph and profile data.
A good start is their list of acquisitions [1] in the space of offline applications, social networking aggregators, contact importers, photo management, private conversations, travel recommendations, check-ins/status updates, etc.
Then consider the FB partnerships with ride hailing services, retail apps, music apps, etc.
Then consider it's own apps and features around events, calendering, photos and instant messaging.
Also consider the state of technology in NLP, AI and image recognition.
Also consider the FB integrations with tens of thousands of apps for social auth, share buttons and advertising.
Writing this post, I have surprised even myself as to how much data they have access to.
There's at least a theoretical possibility that even Facebook does not exactly know why they make certain recommendations. This could be just a machine doing all the work. All kinds of data goes in, predictions are made. Facebook users give feedback by accepting/rejecting the recommendations and algorithms learn and adjust accordingly.
I've just assumed this is through whatsapp. I've logged into whatsapp with my phone number, fb know my phone number (2fa codes), and they've slurped my address book from my phone and then show me connections on fb that are a few degrees of separation. This is the only way I can think that neighbours I barely know show up on the "people you may know" feature -- we're connected through neighbours I know far better.
They probably know your neighbors because you both have the same SSIDs in your vicinity.. I don't see FB not knowing this. Or if you allowed GPS, and neighbors did too, then it's even easier to make the connection.
You don't need to have their number, if someone else has both of your numbers and has uploaded their contacts list then you're linked in that third party's list. Equally likely it's linked via the wifi SSID or GPS?
You are 100% correct about this. People have appeared on "People you may know" when I have absolutely zero connection to them other than I have chatted to the on What's App... I'm afraid I may have to stop using the service soon... was great until FB bought it.
I agree - it is amazing how even a small mention of a product in a whatsapp conversation with my friends turns into a barrage of related ads on Facebook. We have even started teasing each other by mention funny products just to indundate each other with irrelevant ads. So sad - my assumption is: anything you say or do in an electronic device is NOT private.
These kinds of things might be just random. You see probably thousands of ads during the day, but most of them are not registered by your conscious mind. Only when there is a notable connection to something, your brain wakes up and you register the ad.
There's also the option that before bringing up something in conversation, your friend has made some related searches or reacted to a related ad. Facebook could then make the decision to show the ad to you based on your friend past behavior and the fact that you two had a conversation.
> These kinds of things might be just random. You see probably thousands of ads during the day, but most of them are not registered by your conscious mind. Only when there is a notable connection to something, your brain wakes up and you register the ad.
>You are 100% correct about this. People have appeared on "People you may know" when I have absolutely zero connection to them other than I have chatted to the on What's App...
Well, isn't it obvious though? Facebook owns WhatsApp and has access to both "social graphs".
(Plus, it can also infer "people you may know" from people your friends in FB and Whatsapp know).
Also keep in mind that social media/corps don't just use their own data, they also buys search history, credit card transactions, bank information, etc from other companies as well. Granted most of these are "anonymized" but it wouldn't take much effort to link them all together.
For example, if you CC company sends your transaction history to FB without any identification information, FB could still know that transaction history by linking your smartphone location data to the transaction history.
For example, if they know you were at train station at X time and at a Rite Aid at X time and at a McDonald's at X time, they could go "map" that to a "CC customer data tranche". It is highly doubtful that two people would be at same locations at the same exact times with the same transactions at the same time.
FB is big enough to rely on their own data for most of their needs, but selling of your data is big business and everyone is doing it.
And as creepy as FB adding your neighbors to your "people you may know", the real creepy part is the B2B part where corporations sell their profiles of you to each other or to governments/colleges/etc.
And we are just at the beginnings of "datafication" of everyone. It's truly going to be a brave new world.
I would imagine Facebook keeps placeholders for people you know that are not on Facebook. His father had made a contact with one of the relatives not on FB but with the other name. Perhaps the great aunt had him in her contacts as well. A name and/or phone number can represent a person in the graph without being a FB user. Why they made a suggestion so many degrees away is a little curious, but I think they mix those in with the more obvious ones just to check.
TL;DR: I think their graph includes non-FB-using people.
If non-FB/shadow profiles can be derived from phone numbers, or assumed relationships (e.g. "this person probably has parents"), might they also be derived from the contents of private messages?
And this is why under German law the BDSG §34 exists, which allows every citizen to request, once every 12 months, for free, from every company a list of all data the company has on the citizen, how they got it, whom they gave it to, and all statistical models that were derived from the data, or that had an influence on the data.
It’d be interesting to make a BDSG §34 request about "People you may know", just to see how Facebook responds (and if you have to pull them through court)
If a German HN reader would be so kind as to make that request and report back on what they get, I can absolutely positively guarantee that they will be the happy recipient of One Internet Point from me when they submit the story to HN.
You get a CD with every conversation you ever had, every picture you uploaded, everything you shared on that wall. "Deleted" or not, it's on there. Pretty much every interaction you ever had while browsing Facebook.
Cool. If someone figures out the process let's all do it. I quit FB yeaaaaaaaaaaars ago (~2008?) but would be keen to check the shadow profile. I wonder if you can do that? Incidentally I am German by third citizenship.
How does Facebook report inferred connections or connections deduced by looking at a larger graph? Their obligation to tell you what they know about you shouldn't infringe on my privacy.
Do you know anyone who has requested all this information, seen how it was provided and was able to make sense of it, especially from a platform like Facebook? What format is the information provided as? HTML files with graphs? XML files (gasp)?
> It was not a very convincing excuse. Facebook gets people to hand over information about themselves all the time; by what principle would it be unreasonable to sometimes hand some of that information back?
You are not "handing" them anything with an expectation of reciprocity, you (in most cases) willingly give up information about yourself to them in exchange for their free service. They don't need an "excuse" - you were as free to say "no" as they are to you.
I agree that in cases where others give up information about people they know there's no consent from the person (necessarily) but even that is between the person and their peers. There are people in my life who I don't share information with because I know how fast that information will spread if they know it; the kind of privacy etiquette required to respect people's privacy wishes will likely become more commonplace as it's made clear indexing organizations only know what people tell them.
I've been against facebook for some time yet even I don't agree with writers of sensationalist pieces like this. You have what's coming to you when you use third party services like facebook then complain about the imbalance of power they imply. Turns out the reason people who adamantly don't use those services do so other than just 'hipsterism' or whatever you thought were their reasons.
> You are not "handing" them anything with an expectation of reciprocity, you (in most cases) willingly give up information about yourself to them in exchange for their free service. They don't need an "excuse" - you were as free to say "no" as they are to you.
You are not free to say no to facebook.
You do not hand your information to them, they take it. There's a reason they buy information from "data brokers", and it's because it include information you are not willingly giving to them.
Seriously, would love to know. They keep a profile for any person that they think a facebook member knows, so how do I stop them from creating that profile?
I am not used by FB but it most certainly maintains an account of me.
Additionally, handing over data does not give FB full usage rights to it. At least in the EU, data subjects enjoy the basic right of informational self determination. FB may have to answer in a more comprehensive manner with the GDPR coming into effect.
I haven't used my Facebook account in years, ever since they recommended I connected with a person I'd known for 20+ years but with whom I shared no connections or commonalities apart from our nationality. As I remember it, neither of us had listed schools, jobs or interests in our profiles.
Maybe it was a coincidence but it made me very anxious so I left the platform.
It's possible that one of you connected Facebook to a service along the way that provided access to contact information like email or phone number and they used that to make the connection.
Could be, but by and large I've always been a late adopter of social media so apart from a hotmail / messenger account, I think FB was the only common service.
"In practice, the forced revelation of information makes individual privilege and power more important. When everyone has to play with their cards on the table, so to speak, then people who feel like they can be themselves without consequence do so freely -- these generally being people with support groups of like-minded people, and who are neither economically nor physically vulnerable. People who are more vulnerable to consequences use concealment as a method of protection: it makes it possible to speak freely about controversial subjects, or even about any subjects, without fear of harassment."
-- Yonatan Zunger, former chief architect, Google+
Or it is just a chance event? As it is not a controlled experiment and the story concerns a single individual I find the conclusion to be rather shaky.
Napkin math: For just 1 person to intersect: Each of those 10000 people have a (10000/30000000)/100 % chance to get selected, or .00000033%. Multiply that by a thousand selections = 0.00033% chance. Those selections are refreshed maybe 100 times a year so 0.033% chance per year.
There are billions of facebook users, so this should be happening all the time by pure chance. But of course it's not pure chance. Facebook will be selecting from a pool much much smaller than 300 million, and the selections won't be random.
Yep, the article totally ignores it. The fact that she immediately said that they are related, makes me believe she was looking at his profile earlier, and FB just picked up on it.
Facebook is a monolithic threat by inducing the world into accepting the relinquishment of privacy for convenience. The abiding thought that all should remember is it is a marketing (ad) platform. Furthermore, given it limits social outreach to 5000 "friends" (with no such limits on "follows", which still prove difficult to utilize), it is assuredly not intended to facilitate social outreach.
facebook does nothing besides match tons of contact list that idiots upload to their servers when they install any facebook app.
why does people think this is rocket science? even the supposedly clever people here try to imagine complex algorithms. geez. its 1:1 text match on phone numbers.
their genius move was getting enough idiots uploading their entire phone address book for free.
I do agree with your sentiment, but Facebook does a lot more than just that.
They put a lot of effort into tracking sites you browse, their phone app slurps up any possible data to improve your profile, starting with location info, and going all the way to uploading photos from your camera gallery to Facebook, to be analyzed for face recognition.
The other day, I had a couple of friends at my place. At some point, we decided to order pizzas. We did so online, at my usual pizzeria. 30 minutes later (not sure of the delay, but it was before pizzas were delivered), one of my friend check her facebook on her mobile and see ... an ad for that pizzeria. She said she's never seen the ad before, nor ever ordered at that pizzeria.
The best I could guess about it would probably make use of phones gps, detecting we were close and what I was doing on the net, through social share buttons tracking. Anyway, that's outrageous. Facebook use of personal data is that creepy thing we were all afraid Google would do. It really feels like having a glass of wine in your sofa, and the telephone rings and someone unknown says : "cheers!".
Possible, but unlikely: the pizzeria was using a tracking pixel with a purchase event and was targeting ads to friends of purchasers. I don't think you can do this exact kind of targeting (but not sure) and it's pretty sophisticated for some random pizza place.
By far the most likely scenario is a mix of coincidence and confirmation bias. The pizzeria was just running geo and time of day ads and your friend did not notice or remember until this event made it seem significant.
By the way, a follow up : after that, I installed Disconnect in my browser, removed facebook app from my phone, used a dedicated chrome profile for facebook, installed firefox on my mobile and made it my default browser, and installed disconnect and adblock on it. We'll see if that's enough.
I see all kind of theories about the misterious nature of facebook recommendations. I believe that they use all kind of complex things but most of those unexpected recommendations could be explained by assuming that facebook recommend you people that have being cheking you out. If you saw someone on the street and that person check your facebook, they are recommended to you and it looks like magic because you did nothing. I do not know if this is the case but it is simple and seems to explain a lot of those strange recommendations.
Not entirely relevant to the original story, but Facebook does use machine learning to identify faces as well (similar technology to how Google or Apple uses in their Photo apps).
So even if you've never been tagged in a photo or checked into a location, there is a possibility that just by appearing in a random stranger's Facebook photo, they would be able to place you in the vicinity of said stranger at a particular time, and make more conclusions/connections from there.
Well, if Facebook won't tell you how, you need to go figure out the connection you have with the other N thousand people it recommended to you. You might be able to find the common links, and do some valuable reporting in the process.
If the facebook app can identify unique phones in your vicinity -- IE MAC assesses, then it could form links without facebook being installed on the middle devices.
Funnily enough, I'm not FB friends with my wife, and she never comes up on my "People You May Know". I spend more time with her than anyone, so maybe FB's algorithm isn't so great?
I don’t except for signing on to other sites (its easier than creating and remembering a password) and for Wi-fi in Barbados (I’m not paying their mobile data rates).
So do we think Facebook and Amazon are just figuring out every relation between things in the world and they'll work out how to monetize it later? Or, what exactly is the hypothesis here? I have seen them accused of some pretty impressive leaps of inference lately.
> I told the spokesperson that it might be in the search giant’s interest to be more transparent about how this feature works so that users are less creeped out by it.
For contemporary instances, consider that the present U.S. administration openly supports racist, neo-Nazi, White Supremecist, and similar organisations. Social-graph association of undocumented immigrants, Muslims, gang- (or non-gang-) members, or similar inferences, all come readily to mind.
There already is a US example. Virginia instituted a "one-drop" law in the 20's and scoured through old census records to label many mixed race people as black, subjecting then to Jim Crow laws.
Add to that accusations that the FBI used census data to build maps of Muslim communities post-9/11. We like to pretend that since such uses of census data are 'illegal' right now, that they won't suddenly become legal on the next national security whim.
Those are in fact the most relevant elements of this question. They are elements which, as in the case of the Netherlands, don't matter at all, until they mean everything.
Accusing the current administration of openly supporting neo-nazi groups is political flamebait.
You are certainly entitled to your opinion on how "serious" the situation is but I, (and probably others) would appreciate it if you kept the hyperbole and rhetoric minimized. Your original point stands fine without it, and in fact that accusation weakens your entire post and makes it appear little more than a simple troll.
Hold on. That article says that Trump claimed there was "blame on both sides", followed by open repudiation of neo-Nazis and the KKK as "criminals and thugs". I'm not being intentionally obtuse: failure to condemn neo-Nazis and white supremacists is a pretty huge failing for a president (though no worse than what I expect of Donald Trump), and those who condemned his remarks were absolutely right.
But words have meaning, and calling this "open support" is flat lying. He's unbelievably terrible at the job of president and human being: you don't need to lie in order to make the case that he is.
Just because your bubble (and my own) consists entirely of people who are aware of how terrible he is doesn't mean everyone is, and lying for political reasons is exactly the kind of thing that makes big chunks of the country inhabit alternate realities where all news is fake and the notion of a fact disappears from political discourse. You should be ashamed of yourself.
Trump pardoned someone who bragged about his self-described "concentration camps" and policing practices targeting people of Latin American descent, then bragged about how he wasn't going to listen to judges who told him to stop doing it. (Among literally hundreds of other instances of corruption, cover-ups, civil rights violations, and humanitarian abuses.)
I don't know how much more explicit you expect his support of white supremacists to be. (Honestly, I doubt you're engaging genuinely at all.)
> I don't know how much more explicit you expect his support of white supremacists to be.
Dude, are you kidding me? You're looking at someone who said:
“Racism is evil and those who cause violence in its name are criminals and thugs, including the KKK, neo-Nazis, white supremacists, and other hate groups that are repugnant to everything we hold dear as Americans,”
and you can't think of a way for him to support them less? Are you really that obtuse, or are you just playing dumb?
Here's a really simple, utterly obvious way he could have been closer to "open support" (other than uh, openly expressing support......): _not repudiating them_. I honestly can't believe I just sincerely explained a tautology to someone.
Neither article of which suggests in any way what you posted. Your opinion, and your feelings are not the "truth". I understand you have strong feelings, but please, stop polluting this space.
> the present U.S. administration openly supports racist, neo-Nazi, White Supremecist, and similar organisations
I try not to follow what the guy says too closely, but do you have any source for this? The last thing on the topic I recall him saying was a repudiation of said groups.
I feel like I would've heard about it if he openly supported any of these groups.
While I can see how someone could draw similarities in the aggregation of user data, I disagree about that being a comparable system for many reasons. Namely that they aren't government systems, and the political environment doesn't reflect, nor are projected to reflect a state of genocide.
This falls into the same line of thinking as "I need my assault weapon to fight the government when it becomes tyrannical." It's an unreasonable conclusion that comes from a reasonable set of first principals because the landscape is so different now than in the comparing period.
It's not necessary for data systems to belong to governments, or for them to be accessed by the local government jurisdiction.
For examples of each:
AT&T maintain call-level history of its subscribers dating to the 1980s, which has been provided to the US DEA.
Google have disclosed information to both Chinese and Russian nationals, through extralegal, non-discretionary disclosures. The first occured when the accounts of Chinese nationals involved in the Free Tibet movement were hacked. The latter when an email account belonging to Hillary Clinton's campaign chairman, John Podesta, was hacked by parties strongly linked to the present Russian government.
"You can't disclose what you don't have" is a well-established principle of data security. You'll find variants of it offered by Gene Spafford, Bruce Schneier, Eben Moglen, Cory Doctorow, and others. Or in the alternative, "data are vulnerability".
Whether you listen to the left or the right, everyone agrees the landscape is, if at all different, worse than ever, and requiring assault weapons to defend from government tyranny. Literally, the left is calling the right Nazis, and the right is calling the left Nazis. Would you prefer the citizenry is unarmed against the Nazis?
If you're in the center, like me, then you recognise there are credible reasons the landscape is similar enough. Any nation can turn fairly quickly as the political tides shift. History is replete with examples. It is usually only a matter of months.
Separately I'm getting all kinds of people from that town, I can only imagine why, but it's pretty damn scary. And of course I have to wonder who I'm being recommended to and/or my family is being recommended to.