Facial recognition software is just nothing but good vibes, right?
If you use social media, have a drivers license, shop in stores, and walk in public, chances are good that your faceprint will soon be assigned to your identity, and eventually be used on a daily basis to build a profile of you at a level of detail you hoped would never be possible. Facial recognition technology is primed to hit the big time, and is already allowing everyone from small start-ups to a multi-agency US government program to collect information about where you go and what you do. The end result, at least as it looks right now, is loosely regulated biometrics databases that threaten the very concept of anonymity.
You may have heard of facial recognition technology (FRT) lately in the context of identifying suspects from CCTV footage—such as the 2011 riots in London and Vancouver, 2009 protests in Iran, or even the lack of its use as seen in the hunt for the Boston Bombing suspects. You may also know it from the moderately functional security mechanism on your Android phone, or perhaps it immediately makes you think of the Xbox One and its required Kinect camera that will sit there watching your living room, waiting to identify everyone who sits on your couch.
FRT has finally gotten advanced enough that it makes sense to start integrating it into everything from national infrastructure to consumer products. There are still a few kinks to work out—things like bad lighting, indirect images, and obscured facial features can still fool the tech out there now—but there are enough people hammering away hard enough at these issues that its ubiquity is inevitable.
Besides the challenge of perfecting the technology, there is also the hurdle of public perception and acceptance. Google CEO Eric Schmidt said way back in 2011 that “we built that technology and we withheld it. […] I’m very concerned personally about the union of mobile tracking and face recognition.” It is not often that you hear the head of a company that has more money than God from creating targeted advertising based on your emails and browsing history say that he is uncomfortable with technology that helps him know more about you.
Google seems to be strongly maintaining this position, as evidenced by its statement telling everyone not to make Facial Recognition apps for Google Glass a few months ago. The company is not entirely out of the game, though.
Google acquired facial recognition software company PittPatt in 2011, and filed last month to use unique facial expressions to unlock mobile devices. Considering Google's acquisition and patent history, both moves may be attempts to prevent competitors from gaining access to the tech, but the firm appears interested in face recognition at a basic level, at the very least for Android gesture-based controls and biometric passwords.
Many companies are working on facial recognition tech for law enforcement and DMVs, but perhaps the most progress is being done by tech giants for consumer products.
Of course, with the huge possibilities and open platform of Google Glass, other developers can get into the FRT game on their own. That includes Lambda Labs’ recently released beta version of a facial recognition app for Glass, made possible with the help of a homebrew operating system, yet another first in a long line of unintended Glass firsts to come.
There are indeed many companies working specifically on facial recognition technologies and making them available to clients such as law enforcement and DMVs, but perhaps the most progress is being done behind the scenes of tech giants to be included in consumer products.
Even if you are not on the Google train, you might be on Facebook, which acquired Face.com last year. Facebook is trying to improve its facial recognition algorithms so that it can suggest names of people to be tagged in a picture even more accurately than Google, and that research is sure to help reduce the influence of pose, occlusion, and resolution that is hindering FRT. Also of note, Facebook currently has the largest dataset of faces, with around 1.15 billion monthly active users, many of whom who do use their actual face in their profiles.
It's easy to get spooked by government-run FRT efforts, but it's fascinating to think that the technology is being driven forward by companies focusing on consumer tech, such as for better cosmetic consulting.
Apple is also putting in work, though its stakes are higher because it envisions FRT as a necessity for more security in mobile payments. Not only did Apple acquire the behemoth fingerprint reader manufacturer AuthenTec last year, it has also been hammering away at research of their own to improve facial recognition in real-time even in off lighting and blurry/moving images.
Microsoft is in the game as well. The new Kinect system improves upon its popular predecessor in its ability to recognize individual users by face, detect emotions from expressions, and create skeletal models of movements as precise as the flex of a finger. Microsoft is poised to continue to make advancements in the FRT game thanks in large part to their dedicated lab in Israel and the software such as OneVision they are working on.
Heaps of smaller companies are developing more unique ways of using FRT that will probably be bought up by the big guys sooner or later. Affectiva “reads facial expressions to measure the emotional connection people have with advertising, brands, and media.” Disney’s new cruise ship has interactive art walls spread out all over the ship that recognize your face so they never show you the same thing twice. Vesalis provides highly accurate FRT for makeup counters to demonstrate how beauty and hair products will look. Facedeals allows you to check in with your face when you go to a bar or restaurant for personalized coupons based on your social media history.
Of course, it is easy to focus on video games and futuristic makeup counters, but they're only a part of the larger the debate we are attempting to have in America about privacy versus security. As the war on terror drags on, we've come to realize just how far we have let the government infringe on our personal lives in the name of homeland security. Last month, Barack Obama said that “we don’t have to sacrifice our freedom in order to achieve security. That is a false choice.”
But the debate isn’t really about giving up freedom, at least not in the way politicians would frame the word. It is about losing privacy, which appears to be a far more abstract freedom to folks in Washington than, say, those protected by the First and Second Amendments. It can be easy for freedom and privacy to be interchanged in heated debates and media volleys, but just because I am free to go about my day does not mean I am not being watched every step of the way.
Visual computing and HUDs for police use have been in development for ages, and are close to being field ready. Integration of facial recognition tech could allow officers to do something similar to license plate scanning, but with people.
Should we be concerned? There is currently scant evidence of the government abusing the information it is collecting, and it’s not like we haven’t already surrendered massive amounts of personal data to private entities that have created profiles of us for their financial gain. Is it so wrong that a government should have the same sort of database in the name of security and organization?
Just because they can, doesn’t mean they should, as evidenced by other countries’ recent forays down this path. Last year, the French National Assembly passed a law calling for a new biometric ID card for its entire population. Over 45 million individuals would have their finger and face prints digitized and stored in a massive centralized database. This initiative was immediately denounced by many, including 200 members of the French parliament who referred it to the Conseil Constitutionnel on the basis that it is incompatible with European citizens’ fundamental rights to privacy and the presumption of innocence, among others. French senator François Pillet summarized this controversy well when he referred to the bill as a time bomb for civil liberties.
The United Kingdom has been building a “National DNA Database” since 1995, and has, as of March, 2012, an estimated 5,950,612 individuals on file. Of course this has helped solve heaps of crimes, but it also puts your unique biometric information at risk of being used and distributed without your consent.
Genetic research has been done using the database, commercial companies that do research have DNA information stored on their own computers, and in 2007, five civil servants were caught stealing DNA information to establish a rival firm. Shami Chakrabarti, director of civil liberties organization Liberty, referred to this database and a proposal to expand it to every man, woman, and child in the UK as a chilling proposal, ripe for indignity, error, and abuse.
We obviously cannot discuss surveillance culture and biometric identification without considering the case of China. As Mara Hvistendahl explained in Science, the country is on a stated mission to install 30 million security cameras by 2015 (if they haven’t already), or around one for every 45 citizens, at which point the industry is expected to reach 500 billion yuan ($79 billion USD). It should come as no surprise that universities and companies alike in China are developing intelligent video surveillance infrastructure that will help law enforcement search for specific features, such as a man in a red t-shirt, and eventually even specific faces, drastically improving the timeframe of results from the prior system in which officers manually scan footage.
As China’s biometric and image analysis technology improves, so does the mass installation of security cameras and biometric checkpoints within its borders to keep tabs on political dissidents and its population in general. As Nicholas Bequelin, a researcher at Human Rights Watch in Hong Kong, told Hvistendahl, the implementation and use of these technologies needs to be monitored “by a strong civil society or a political system that is responsive to citizens’ concerns.” In China, however, this “is just not there.”
The stated future goal of the FBI is to identify subjects in “public datasets," based on publicly available photographs from sources like Facebook and the Internet at large.
These questionable scenarios abroad aren’t stopping the trend from making serious headway in the US, but so far FRT hasn’t been applied in a way that substantially impacts your daily routine. You might see some changes soon, however, as your unique biometric information is slowly being amalgamated into a national database.
The FBI has been working on the Next Generation Identification program with the help of its friends at Lockheed Martin, IBM, Accenture, and BAE Systems, among others. The NGI system is a biometric database that could be deployed as early as 2014. This $1 billion database, which would be the biggest of its kind in the world, will include DNA, iris, voice, and facial recognition data in addition to more commonplace biometric information. The photos to be included will come from a variety of sources from DMV databases to private security cameras.
The stated future goal of the FBI is to allow law enforcement agencies to identify subjects in “public datasets," based on publicly available photographs from sources like Facebook and the Internet at large. The FBI is already being sued over the lack of transparency of this project, which are now joined by lawsuits against the NSA.
The NGI program should come as no surprise based on other government programs, such as Stellar Wind and PRISM, that have come to light, and is just another unfortunate example of how the US government is keeping tabs on all of its citizens, even the innocent ones. As Electronic Frontier Foundation attorney Jennifer Lynch explained in a phone call, “I think we’ve seen from oppressive regimes in other countries and even from FBI’s own practices in this country under Hoover that the government can take that information and use it against people in ways that we might not be able to anticipate right now. That’s why we need to limit collection programs before they even start.”
It isn’t just the spooks that are coming after your biometric information, however. The new Senate Immigration Bill, buried deep within its 800 pages, calls for a centralized database of every legal US worker or potential worker. This new evolution of E-Verify seeks to implement a biometric verification system for all working Americans. In other words: No picture, no job, sir.
As Stellar Wind whistleblower William Binney said in a New York Times video last year, “just because we are a democracy doesn’t mean we will stay that way.” We may have many freedoms under our government, but we are losing our democratic society as we lose the principle of ultimate control over the government by the people that the country was founded on. It is disconcerting to realize that we are struggling with the same kind of surveillance concerns as a population as the Chinese, whose government is not famous for an impeccable civil rights record.
If we want to maintain some privacy as our government struggles for homeland security, we are going to have to fight for it ourselves. Even the Fourth Amendment, which protects against unreasonable searches and seizures and is generally thought to suggest privacy, might not apply to some of our biometric information.
In a recent Supreme Court opinion, Justice Scalia indicated in his defense that taking a photograph of someone would not implicate the Fourth Amendment. As Lynch explained, he “[argued that] it wasn’t a search or seizure because it didn’t involve any touching. It’s probably one of those collection practices that will be regulated by statute rather than constitutional principle,” a statute that does not yet exist.
“I think most people truly expect to be anonymous in public,” Lynch said. According to her, anonymity is a crucial element of public spaces and communication, and is what allows us to freely participate in a democratic society.
“[When people] are being watched,” she said, “it chills their behavior—it chills who they will associate with, how they will act, and what they will say. That’s not a society I want to live in or raise my son in.”
It is chilling indeed to think about how we are increasingly going to be identified by an algorithmic representation of our faces instead of simply by our names. The technology to identify us is poised to be cheap and effective enough to be used in devices as simple as highway toll transponders. It won’t be long before law enforcement uses it as regularly as sirens and handcuffs, or everyday devices require your faceprint to be accessed.
FRT gets the 60 Minutes treatment. It's stunning where the cutting edge lies.
The current hurdle to widespread FRT adoption is that it's still pretty difficult to recognize faces from every angle and every lighting condition. But with facial recognition being worked on by just about every major tech company, as well as the government, those hurdles will be passed soon. And initially, FRT integration will likely be optional and sometimes pretty cool—or at least harmless. With a Kinect camera in your living room and your smartphone camera in the palm of your hand wherever you go, it makes sense for tech companies to create very accurate models of your face in the name of security and interfacing.
At the very least, being able to skip movies on Netflix by blinking at your TV is another feature to advertise . And it's easy to imagine a time when the convenience of using these identifiers will far outweigh the hassle of proving your identity in other ways, much like the plethora of sites that encourage you to log in with Facebook.
Like it or not, your face is the next frontier of identification. This is the beginning of the end of anonymity, as your faceprint increasingly becomes tied to more things like your Google account, license plate, SSN, and an endless list of other identifiers.
Without clear legislation about how biometric information is collected and used, we are soon going to be hit with the realization that it is becoming impossible to be anonymous. “The problem with FRT or any kind of identification system,” said Lynch, “is that it has the ability to identify people and remember them years down the road.”
Currently there is only legislation to limit the use of biometric technology in three states—Illinois, Texas, and Washington—and nothing is on the table at the federal level. In Illinois's case, the law calls for clear explanations of what biometric information is being collected and how long it will be kept for, and in the case of a security breach, your face is worth $1,000 if due to negligence and $5,000 if intentionally or recklessly misused, unless damages are shown to be greater than that. Texas has similar stipulations, although your identity is evidently slightly more valuable, as violators can be fined up to $25,000 per incident. Of note is that these laws do not apply to government agencies or their contractors.
“It really seems like our government is overreaching and not doing it all it can to protect our privacy,” explained Lynch. Between laws that permit increased surveillance and a lack of laws to control the use of biometric technology, there are very few limits on how much information will be collected about you, who will store it, and what it will be used for.