Is Facebook's Facial Recognition That Scary?
from the how-so? dept
There has been some concern here and there about Facebook’s face recognition technology, though I will admit that I still don’t see it as being that concerning. It only works on people who you’re already listed as being “friends” with — not just randomly identifying people in crowds or anything. As such, it seems more like a handy tool for better providing info in images. Some have raised issues about Facebook opening it up beyond just friends, or potentially exposing ID info in photos to others, like advertisers. And while Facebook has a reputation for changing its privacy terms in ways that freak people out — and which possibly expose info that people originally thought was private — I’m still missing why the face recognition is so problematic. It’s no different than what a user would/could do right now — identifying an individual in a photo and tagging them. If facial recognition is a concern, why isn’t that (human) action just as concerning? Both involve connecting an image to a name.
Perhaps I’m missing something — and I fully expect readers here to clarify exactly what it is in the comments.
It certainly seems like governments around the globe also think that there are problems with the facial recognition — and they may act on it. Here in the US, Senator Franken quizzed a Facebook exec on this issue. He properly noted that Facebook doesn’t do a particularly good job disclosing that the facial recognition technology is there — and that’s something that could be improved. But is that a legal violation?
The bigger issue, though, may be over in Europe where there are a variety of investigations of the technology going on — with Ireland and Norway leading the way and expressing concerns. Europe, of course, takes privacy rights against companies to a different level than we do in the US, though that sometimes leads to wacky, nonsensical outcomes, like German residents being able to blur their houses on Google Street View. But of course, the fact that anyone driving down the street can see what’s there is ignored. It feels somewhat similar with facial recognition. If you’re in a picture, you’re in a picture. A human can identify you. Is there a major difference if a computer can also identify you?
Filed Under: face recognition, privacy
Comments on “Is Facebook's Facial Recognition That Scary?”
I still don't see it as being that concerning
Mike you are so trusting. How long till the NSA/CIA/FBI gets an office at Facebook?
Re: I still don't see it as being that concerning
First rule of Facebook: Never talk about the fact that the NSA/CIA/FBI already has a secret virtual office at Facebook (in the cloud, with all the warrantless wiretapping they can eat).
Yes
Is there a major difference if a computer can also identify you?
It’s a difference of scale, so yes. Much like public records can be problematic once they’re machine-searchable.
Re: Yes
Yes, this. The aggregation of lots of individually meaningless data points results in a massive invasion of privacy that almost everyone would object to if it happened with a single surveillance action.
Re: Yes
Of course, this is also dependent on people continuing to ignore the options available to them. While it is true that the facial recognition enables the creation of a large amount of machine searchable data, a few minutes of maintenance can eliminate all of that.
Most people are completely unaware that they can turn on “tagging review”. This means that if anyone in your friends list tags you in something, you have the choice to accept or reject the tag before it goes “public”. The picture is still up, but the tag won’t display. By default, this feature is disabled. Oh, and if someone NOT in your friends list tags you, you’re supposed to automatically be given the accept/reject option.
On a side note, Facebook would probably get a better response if they moved the functionality to a button on the photo called “AutoTag”. Instead of automatically applying it, force people to intentionally run the facial recognition.
/much ado about nothing
Re: Re: Yes
While I’m very sympathetic to individualist responsibility arguments (most definitely in this case, so don’t take this the wrong way), the problem with that type of thinking is that it tends to create an inverted pyramid of responsibility. We live in a highly specialized society, and one where most people use multiple computers multiple times a day, but have exactly zero idea how to protect it, themselves, and their data while using it.
If you take every position where a sort of “caveat emptor” argument has been used to suggest that people ought to be watching out for themselves, you see that the people on the bottom are expected to be super-paranoid experts in just about everything, and the movers and shakers have very little responsibility restraining virtually anything they do.
Re: Re: Re: Yes
Perhaps it’s my cynicism poking through from having far too many friends and family members constantly ask me for help with their computers, but I’ve always believed that if your going to use a tool, you should at least be somewhat knowledgeable about how it works. That should include how to keep it running properly. The same would apply to cars, power tools, home ownership, cell phones, etc. With computers, the first thing I do with any new site or application is to run through the options/settings/preferences and see what is available.
You wouldn’t leave your car stereo presets on the settings from the factory, so why leave the default settings on your Facebook account?
Re: Re: Re:2 Yes
My power tools don’t add new features to themselves every time I put them Mike.
Your expectation is unrealistic. You can tell this from how very common it is for ordinary human beings with ordinary levels of intelligence or higher, failng to meet it, despite not being particularly irresponsible examples of human beings.
You could look at the human race and say a sizable chunk of it is just not good enough for Facebook or other technology that is routinely not adequately understood by a sizable chunk of its users, but I’m not sure what advantage or utilitly just dismissing a huge chunk of humanity as inept is supposed to offer us.
I know it’s fun to look down our noses and judge folk, but the fact is people are people are people and people don’t exist for the purpose of serving Facebook or other examples of technology.
Re: Re: Re:3 Stupid people
Stupid people deserve to get screwed over until they clue in…
Really?
I’m surprised this is such a difficult question to wrangle with for a self-described privacy advocate. I should say now that I’m thinking of the issue generally, and not just as specifically related to Facebook, but that’s only fair since I haven’t used Facebook in a number of years and I can’t really comment on that feature or how it impacts a Facebook user.
The concern in — well, probably all of these cases is that while, yes, a human can perform the function described, a computer can provide it much faster, in many more situations, and can process and store the results in a much more far-reaching manner than is normally possible. With respect to the blurred German street view, yes, anyone near their neighborhood can drive by, which if they live in a big town is a pool of maybe 1000 people they sort of have a connection with (living nearby) and even though they’re strangers feel somewhat more comfortable with. That’s a big difference from the entire population of 7 billion people on Earth being able to log onto their computer at any given moment and scope out your house.
It seems to me like your reasoning could be applied equally validly to someone objecting to deep packet inspection or the like. “Well, couldn’t someone just stand over your shoulder and see what you’re doing online?” Well, yeah, but that’s someone you know and someone you can monitor (accountability!), and they can’t simultaneously do that for millions of other users, remembering and executing every action with perfect clarity, and compiling the results for any number of parties (intended or not) that have purposes you would never explicitly authorize if given the chance.
And that’s the real problem in every one of these cases — you never get that chance.
Re: Really?
You can’t really monitor people walking or driving down the street looking at your house either. Should you cover your house in sheets to keep them from seeing it?
Nothing to Hide!
What are you hiding?
What are you afraid of?
Criminal scum!
Governments Fail, though maybe not a bad reason.
US congress grilled Facebook and asked exactly how Facebook’s facial recognition worked. I’m slightly astonished, but not surprised that Congress did ask about its capabilities. Honestly, the first thing that came to my mind is that it would be used in targeted advertising. Congress was worried about its ability to recognize faces in the CIA and asked if it automatically identified a user without tagging. I recently tried the feature with my wife, and it basically makes tagging easier because it picks out the face much like a camera does in autofocus. So it’s not scarry at all 🙂
Re: Governments Fail, though maybe not a bad reason.
“US congress grilled Facebook and asked exactly how Facebook’s facial recognition worked. I’m slightly astonished, but not surprised that Congress did ask about its capabilities.”
Probably because they have a vested interest in using the technology themselves. The government has a disturbing fixation with any technology which can be used to monitor, snoop and index citizens.
In that regard, I don’t trust Facebook either. Why should I?
Re: Re: Governments Fail, though maybe not a bad reason.
True to a point. But I have my Facebook privacy settings set to “Friends Only” and made sure my photo settings are set to “Only friends can see photos of me”. The hugely common misconception about FaceBook’s privacy issues is that it’s Facebook itself, but when you look closely, you aren’t getting spammed by adverts in your e-mail on the page.
No, the real problem is some of the apps you run. Normally, if you have Facebook set to Friends Only, none of your photos show up on Google Cache. Case in point, my mother in-law ran an app called Mafia Wars, the app and all I had to do was look up her name to get her profile picture up. She since then deleted the app and since then, I cannot use Google search to see her profile picture.
About that article, here’s my response:
Depends on the context…
Don’t forget, a computer can do “facial recognition” much faster and many more times than a human, with more errors. So, context is important.
That, and image recognition technology of any variety is not perfect or error-free (OCR comes pretty close in some areas).
Re: Re:
Then, the inevitable day comes when Content ID Facebook facial recognition software flags your face for automatic takedown because you happen to look like some famous person, and you obviously don’t have the publicity right to look like the face that nature gave you. Next comes the court ordered mandatory ski mask while going out in public or in lieu of that facial plastic surgery, so you won’t be infringing on that famous persons likeness that is a wholly owned subsidiary of whoever was the highest bidder that day on FaceMarket.
I know, it sounds like science fiction now, but then so do a lot of the crazy laws and just as crazy lawsuit fillings these days.
Usage Creep
“Some have raised issues about Facebook opening it up beyond just friends, or potentially exposing ID info in photos to others, like advertisers.”
I think this is issue for a lot of people. Facebook has shown again and again that they can’t be trusted to keep private what users have asked be kept private. Even if photos aren’t openly auto-tagged except with friends at first, how long will that last? Do you really think that Facebook is not going to fully exploit identifying everyone they can in every photo they get? “You’ve been recognized in photos with these three people. Why not friend them? Even if you don’t, we’re going to assume you know each other and use the connection to target ads, etc.”
Where does it stop?
Pretty soon FR will be incorporated in to the surveillance grid that is being set up so every move is tracked. Then there is the tag tracking software the will follow your car on the freeway and every other road that has cameras. They will track your financial records. They follow you through your phone. When do we stop it? When will enough be enough?
I am not a farm animal to be tagged and tracked.
Re: Where does it stop?
Actually humans, so far as I have been able to identify, are the only species to brand and then (socially) track themselves. Facebook wasn’t even a glimmer of an idea before this was practiced.
It is far too late to stop it. The war was lost with the first pair of Levi’s bought because of the stitching and little red tag. Sorry.
Example
Say I have two hobbies, with two different groups of friends. One of those hobbies might cause me a great deal of embarassment with the other groups of friends. Not harm, but I’d just really rather it didn’t happen.
At the moment I can simply ask the friends in hobby group A not to tag me in photos and that’s it. Job done.
Now I am concerned that people from hobby group B will see photos of me in hobby A and that might make things very uncomfortable.
“If facial recognition is a concern, why isn’t that (human) action just as concerning?”
The only thing I can think of is that there’s some personal human knowledge that gets in the way of the action. For example, you have a picture of a workmate at a party where he’s dancing with a few other people. One of those other people is a girl he had a brief fling a couple of years ago, his current girlfriend hates her and tagging him would open up all sorts of problems. She’s not friends with the person who uploaded the photo, so won’t see it unless it’s tagged…
In other words, while both the machine and human scanning the photo would recognise the face, the implications of doing something with that photo is something the human would be far more aware of than the machine, and thus would be less problematic.
On top of that, the machine might be able to build connections that are damaging or wider reaching than people would like. For example, at the same party you only really recognised 20 or so people there, and only have a real personal relationship with a handful. The machine wouldn’t know which and build information about personal relationships with everybody at that party – even people you’ve never met – information which could be problematic if exposed.
That’s my take, anyway. In terms of general usage of the information as it is now, I don’t see the problem, though I think most people fearing this are doing so due to potential future abuse, not the current usage.
The issue I have is not with facebook using it to tag photos. The issue I have is that they are building such a database. They will hold a data base with all the recognition data. This could then later be taken and used for other uses.
As for why this is different than people using it. Well the government could not pay hundreds of people to watch video feeds looking for people. It is very possible to have a computer analyze video feed picking faces out of groups once you have all the recognition data.
So is what they are doing now a worry for me? No, not really, But I am worried about the potential for abuse in the future. It is far to easy for someone later to say “we could have stopped this murderer if we had a system to spot his face and find him sooner” or “It would have helped us catch him quicker if we could have ID him from the video by comparing it to this data base”
Law of unintended consequences
Here’s the scenario (and this is out at the depths of bad things, but so be it) that I find potentially chilling about this sort of thing.
I take a photo at a public event of a few of my friends and random people in the background. Upload that photo to Facebook and possibly tag my friends (or not, it doesn’t matter). In the background, Facebook’s facial recognition is not just trying to find my friends in the picture, but match any face. It turns out that someone visible in the picture who I have no connection to is a wanted fugitive and the FR program identifies this and informs the FBI/NSA/Army Intelligence/Whoever.
In the best case scenario, someone calls me and asks me about it and I say I don’t know who that is. In the worst case, I’m detained for “National Security Reasons” and questioned. Given that opening, the unnamed agency gets a warrant to sift through my online life and decides because I complained about (TSA, IRS, NSA, Obama, Bush, DEA, ATF, etc) that I may be lying about my association. Longer detainment. More questioning.
It’s the machine automation of such things that causes problems. The government rarely apologizes when they break your door down for being a false-positive. Post-911, I tend to believe we’re only one bomb away from being a militarized state with 0 civil liberties, but I’m a cynic.
It would be much better if facebook created the tags as drafts and only attached them to the photos for others to see if and when the account owners approved them.
Facial recognition impact in Iceland
One of the thing that was half-funny, half-horrible that Iceland experienced with Google’s face-recognition technology was that it would take a family photo and label the children of husbands who had been cuckolded as their “real” fathers (family resemblance in Iceland is usually VERY strong). There has been a long history of completely ignoring these situations in Iceland and bringing the kids up as though they were the true son of the father. However, when a label is applied to a photo for the whole family/village to see difficulties can arise.
Re: Facial recognition impact in Iceland
I’m sure that Google will say something like “It’s for the children” to help keep the accidental inbreeding down, because now not only can you check a name but you can also check a face too.
Time. com – Icelanders Avoid Inbreeding Through Online Database
February 9, 2012
Nowadays, some light Internet stalking is as common a pre-date ritual as showering or putting on a clean shirt. But for Icelanders, that online screening process can include running a date?s name through a genealogical database.
Sound ridiculous? Consider this: when you live in an isolated nation with a population roughly the size of Pittsburgh, accidentally lusting after a cousin is an all-too-real possibility. But a search engine called ?slendingab?k (the Book of Icelanders) allows users to plug in their own name alongside that of a prospective mate, determining any familial overlap. The site claims to track 1,200 years of genealogical information about the island?s inhabitants. Anyone with an Icelandic ID number ? that is, citizens and legal residents ? is accounted for, the New York Daily News reports.
Not only can the site rule out courtships that might be a bit too close for comfort, but also it helps users determine if they share family ties with any famous Icelanders. One man learned he was related to the singer Bj?rk seven generations back, as well as Prime Minister J?hanna Sigur?ard?ttir, Global Post reports. He also discovered that his ex-wife was his seventh cousin.
The venture resulted from a collaboration between a genetics research company and a software entrepreneur. If only the site could also rule out close-talkers, all the world?s dating problems would be solved.
Isn’t the facial recognition always there, under the surface, regardless of if it’s publicly displayed with a tag that says ‘this is you?’ I would think the problem is that “it only works on people who you’re already listed as being “friends” with” is a limitation in the interface Facebook presents to ordinary users. What about the interface they present to themselves or to law enforcement? It’s not a limitation in the tech, that’s for sure.
It’s a pretty easy thing to understand Mike. Apparently you don’t deal with people away from the computer much (AFK for you).
Basically, I can have a friend take a bunch of pictures at a party, whatever, and say “Don’t tag me, I don’t want anyone to find this picture of me with the giant bong”. They respect that, and as a result,, searches don’t turn it up, and there is no harm, no foul. Remember, the other people who were at the party (and likely to see the pictures online as friends) already know I had the bong. No big deal. But I don’t want my future investors to see it.
Now, facial recognition would make it that Facebook would tag me anyway. I can’t argue with the machine.
It’s called dealing with real people, not machines. You should try it sometimes!
uhm..
The fear is that they have the largest database of facial data on the planet, all directly tied to names. They have the data points mapped and saved. It has your name attached. Right now, it’s private, and presumably not being shared or sold. Right now.
Now I’m sure you saw the #TrapWire stuff yesterday. It’s already running in NY, DC, and lord knows where else. Does anyone really think the gov is not looking for ways to plug that Facebook data into the #TrapWire system? We did all the ground work. We tagged photos to provide a great database of photos attached to our names. Algorithms have all the data they need to increase the accuracy of correctly identifying these images.
What could go wrong?
Because other people put my picture on sites where I am not a member
I care. I am not a member of Facebook. I choose not to share my personal details with the world because you do not NEED to know that information. You have no right to it. [And I am not a mass murderer. 😉 ]
My picture is on Facebook. Family members have put it there. I would rather not have it there. But the network effect of data analysis on the Internet, and the ability to put together widely separated bits of information to reveal something that was never intended to be revealed — now that is creepy!
Recently, I went through my links to support privacy in an organization that I belong to. I fished some out for you. I note that some of them are from Techdirt. So here they are:
1. A good article on why we should be protecting anonymous speech that looks at some of the things that
make it hard to speak anonymously. This lady has a reason to be anonymous. Her life and her family
was threatened when she published material that challenged false statements:
In Defense of Anonymous Speech ~pj
http://www.groklaw.net/article.php?story=20120212133227775
1. People who care about privacy and anonymity:
New York Greater Metropolitan Area chapter of the Internet Society
ISOC-NY Event: Eben Moglen “Freedom in the Cloud” – 2/5/2010
http://isoc-ny.org/?p=1338
Everyone wants a piece of you these days: Google, Facebook, Flickr, Apple, AT&T, Bing. They’ll give
you free e-mail, free photo storage, free web hosting, even a free date. They just want to listen in. And
you can’t wait to let them. They’ll store your stuff, they’ll organize your photos, they’ll keep track of
your appointments, as long as they can watch. It all goes into the “Cloud”.
3. A ‘Trustworthy’ Social Network For The Occupy Movement: Even If They Build It, Can They
Ever Trust It?
http://www.techdirt.com/articles/20120106/09421517302/trustworthy-social-network-occupy-
movement-even-if-they-build-it-can-they-ever-trust-it.shtml
4. What’s In A Name: The Importance Of Pseudonymity & The Dangers Of Requiring ‘Real
Names’
http://www.techdirt.com/articles/20110805/14103715409/whats-name-importance-pseudonymity-
dangers-requiring-real-names.shtml
Re: Because other people put my picture on sites where I am not a member
Just a quick reminder….you cannot be tagged unless you’re a FaceBook user.
Re: Re: Because other people put my picture on sites where I am not a member
I accept that is true under CURRENT rules. But Facebook has a long history of changing their rules, either temporarily or permanently, that I have absolutely no faith that it will remain true.
If you have confidence in it, go for it. I have no problem with that. But I choose not to.
The sad fact of the matter is that a distributed network with nobody in charge, where people were able to share their information without losing control of it – has morphed into a huge body of people who are convinced that they are incapable of managing their own data. Corporations are doing it for them, and making vast amounts of money off of them.
Just think – what does Facebook own that is worth the billions of dollars that investors sunk into them since the IPO? It is your data. ISPs have helped foster this by writing contracts with you that don’t match the intent and the technology capability of the Internet. They WANT you to be a passive consumer.
If you are interested in controlling your own part of the Internet, take a look at the Freedom Box [ http://freedomboxfoundation.org/ ]. If you have put a cost on your freedom, then it will look insufficient. But if your freedom is absolutely necessary, then it is priceless, and you will start trying to learn what you need in order to be free.
What we really need, of course, is privacy legislation in the US that isn’t bought and paid for by the corporations.
Re: Re: Re: Because other people put my picture on sites where I am not a member
Im not worried about the rules changing until they actually change any rules on it, sountil then it’s pure speculation as to what changes.
Now that being said, I think that’s why my government actually grilled them, making sure the features did stay private. I recall seeing it in C-Span as a matter of fact.
So really there is no need to worry.
The short short version.
Facebook only recognizes a human face and gives you the option to tag it or not. It only provides names on your friends list so nobody but your friends will see it. Furthermore, it’s the same technology that some modern digital cameras have where they can make a face sharper in your shot. I am not currently worried about it because my Government’s only concern was whether or not the public’s facebook pages would be visible to advertisers and they aren’t.
The privacy implications mirror GPS
A human can tag a face, so can a computer. So what’s the big deal, right?
A human police officer can track you 24/7 by following you around in person. So can a GPS tag attached to your car.
Now do you see the big deal?
Re: The privacy implications mirror GPS
It recognizes human faces only. You have to tag each picture.
The problem is that you are being feed into a system that collects your information and then sells to undisclosed 3rd parties all without your knowledge or consent. If your a facebook user your knowingly (although most remain ignorant) exchanging your personal info for access to a service you find valuable.
Even in absence of the special interface for the FBI and others (I thought that interface was just to stream line request process, not just open data access) you are still only a warrant away from any government agency from having the information.
I will probably be off FB within the next year.I will want a more private life.
Re: Re:
you can never get all the way off, btw.
I “deleted” mine over three years ago. My buddy who is still on it says i simply show up as “inactive,” and that all the photos I had posted that have him tagged are still on there as well. Facebook is permanent. There is no escape. All you can do is leave is quickly as possible to mitigate damage.
Re: Re: Re:
“Mitigate damage” – that you, yourself, caused.
At least you’re owning up to it.
I am not surprised
I am not surprised. Maintaining anonymity and privacy online is becoming increasingly difficult. This is just another step in that direction.
Also, government agencies already have a picture paired with you name and other data. And I am sure their face-recognition software is pretty good as well.
Re: I am not surprised
#TRAPWIRE
It should be opt-in. As EVERYTHING any company introduces in an online service.
Umm…if you haven’t already learned that everything on FB is public…well…you aren’t all that smart.