TSA Plans to Use Face Recognition to Track Americans Through Airports
eff.org
eff.org
Fixed it for you. I know for a fact at least 2 asian countries which are either actively using or trialling this, and not just at airports. Both already have fairly comprehensive facial recognition DBs of their citizens (smile for your driver's license picture!)
I've also heard anecdotal accounts of people who stay too long airside at SIN being questioned by airport police. They walk right up to them - pretty obvious what's going on. My only surprise is that they would reveal their capabilities so obviously, but maybe that's a deliberate strategy ("we know everything so don't try anything").
It's only a matter of time before basically all governments are doing this. The tech is there.
There are many techs out there that are quite dystopian (face scans, mass surveillance, etc.) or outright evil (alleged organ harvesting of death row inmates in China) but yet the governments aren't doing it. Even that article in Polish constitution smells (to me personally) like it came due to secret police back in communism collecting 'dirt' on people, that required 0 tech, just secret informants with small monetary incentives or who were blackmailed with dirt already collected, telling little bits of 'innocent' things and yet I can't imagine many (if any) countries Western democratic doing that.
Why is this necessarily evil?
- it provides an additional financial incentive to execute inmates
Brilliant, Watson! I’ll inform the prime minister’s family immediately.
Well the thing is that is open to interpretation. You can define anything as "necessary".
From what I have heard the typical "foot in the door", if you will, is AML/KYC. There's a convenient intersection there between anti-terrorism (national security) and the everyday (banking). Bring those together and now you have a perfectly plausible excuse to build a biometric database on everyone with a bank account - ie, everyone. And once you've got that DB, wouldn't it be positively remiss not to start using it at the borders?
This is the legal basis of the DB in a certain large asian country you've heard of and, on information and belief, it's being expanded to airports too under the guise of national security. The DB has never been announced or confirmed but its existence can be inferred from KYC banking processes there. The future is now!
> but yet the governments aren't doing it
I would be very surprised if a biometric database doesn't exist in most moderately wealthy countries, even in Poland. They may not (yet) be able to use it legally but knowing the general climate you'd think the various intelligence organisations almost aren't doing their job if they didn't have it.
I know that everything legal is open to really wide and crazy interpretations already but if there weren't such a line to interpret there would be no chance for an anti-scanning interpretation. It could also end up being limited to foreigners since the line protects citizens specifically.
It depends what do you mean by biometrics database, i.e.: Polish passports are 'biometric' and have two index fingerprints stored in them for a few years now. I have no idea if the passport database is accessible to police or if there is any face scanning going on.
I'd say the foot in the door is airports themselves. You already need a passport or ID there so the face scan is literally no new information and only confirming what the papers say. The real creepy tech to me is mass country wide face scanning to track everyone everywhere 24/7 and see what they do, who they associate with, etc. and possible future airport brain scans (like MRI to detect positive or negative emotions in response to a country flag or a picture of state head for example).
All it takes is some demagoguery and opportunism to convince the populace that it is in their interests, like what has happened in the UK and US.
I don't even think it's an organized effort to create a totalitarian state but more how such a state will emerge from the interaction between public, media and politics.
This is why comments such "but spy agencies have always been spying, it's their job" have always pissed me off. The spy agencies will continue to encroach on our rights the more "tech is here". We now have "smart" devices with mics and cameras in every room, so they're going to use that.
If 30 years from now we'll use some kind of "bluetooth headset" that actually interacts directly with our thoughts, then they'll use that, too, to screen out thoughts and arrest you for "extremist thoughts". There's no limit for how far these people are willing to go - just because "the tech is here".
And the sad part is the vast majority of people don't understand new tech, so they aren't even aware what's happening until it's way too late. And then it's already "normalized" and the government agencies cry about "how much they need it" and do slight of hand tricks with "look over here, this tech helped us save a kid from kidnapping!" -- "but don't look over there, where we abused it in 95% of the cases for our own twisted and vengeful interests."
Once we have established that safety is a relative term it becomes pretty clear that ruthless surveillance would yield some safety improvement. I think that you can make a much more meaningful stand against surveillance when you acknowledge that, a opposed to outright denying any possible effect.
In my opinion, that only makes it more likely that this data will be misused in order to prove that the dollars spent yielded _something._ That the cry of "safety" is accepted as somehow obviously true is a significant problem and exacerbates the issue.
- Everyone working for "agencies" do so with no regard for their stated goal and are simply interested in their own personal gain.
- They're genuinely intending to prevent incidents of "crime" and are relatively disinterested in the maintenance of your privacy.
I find the latter more likely than the former, with a mix of the two broadly probable. Given that I find it equally unlikely that they achieve literally nothing. (Frankly even if the former were the truth it seems unlikely they wouldn't prevent something at some point even by accident).
Claiming they do _nothing_ is ill founded. The more valid argument is "But at what cost?"
To be honest, the only way I might even consider that this kind of surveillance is helping would be to see some concrete examples of them literally saving lives.
I do not believe that those performing the surveillance do nothing, on the contrary, I believe they may be responsible for any number of harms (leaking private information, having that information stolen, becoming a party to blackmail, etc.) But I certainly do not believe that they are making any of us safer.
Also, there are lots of actually effective measures to improve safety that don't involve giving governments direct access to what everyone says and does.
I am something like 100x more likely to get struck by lightning _twice_ than I am to be killed by a terrorist
Privacy invasion has to be stopped by law. There is no way, currently, for a person who has public safety as a responsibility to say "no" to this kind of technology.
It's much the same with businesses invading user privacy and using it to drive profits. Without a law forbidding it, everyone has to keep up with the profits of the most "evil" corporation.
We need explicit laws.
I'm not against doing something, I'm against counting doing anything as equivalent to doing something productive.
I completely agree with you.
They get away with that when it comes to gun control, why not with privacy too?
EFF yearly budget 9musd...and they have a complicated history with active political engagement.
ACLU yearly budget 133musd
Both the EFF and ACLU do a lot more than just privacy. The NRA does one thing...and they are really good at it. They also have language in the constitution they can mis-interpret. For privacy we are left with the abstractness of unenumerated rights.
This use of facial recognition is a clear violation of the 4th Amendment. The government is always trying to do shit like this, and it’s our job as citizens to stop them. The big takeaway should not be this pathetic, wimpy “they’re always going to try shit” as if trying were the same thing as succeeding. The big takeaway should be give your money to the ACLU and the EFF.
I feel like you and the parent of your post are misinterpreting what sho said. I didn't read it as him trying to justify it, or even say "the tech is there (so let's just accept it)", I read it as him pointing out that it's not just the TSA (or other US agencies) doing it.
And how do you know that he's a "coward" and not doing anything about it? What are you doing about it, other than presumably giving money to the ACLU or EFF?
> England had a police state when the only way to do that was to open paper mail with steam.
England (I assume you mean Britain) has never been a "police state", however draconian new laws may be in this day and age. What are you talking about?
Some things to specifically consider:
1) 240 years ago, the USA was the first country founded to prioritize liberty. 'Everyone' else had authoritarian monarchs at the time.
2) IMO, 'police' states are fine outside the USA if that is what those citizens want (e.g. Singapore, Australia, etc.).
3) U.K. has way too many public cameras (for me to feel comfortable).
Yes. And so is a human's ability to predict who are good and bad actors in a given situation. The statistical mathematics is very deceiving to most people.
See "Base rate fallacy" for more information: https://en.wikipedia.org/wiki/Base_rate_fallacy
Do. Not. Want.
The problem is that people believe it isn't garbage and give it more credibility than it deserves.
Could automate much of the TSAs job.
I never understood why they don't OOR (optical object recognition) the items going through the xray scanner.
[0] Emotions via Eye Tracking http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.492...
[1] Eulerian Magnification http://people.csail.mit.edu/mrub/vidmag/
Why on earth every single time the US government does something that is, err regretful, there's always a bunch of white nights ready to point out something of the sort 'but all the other kids do it'? Sounds like a recipe for societal disaster indeed.
What we should we at the very least considering is whether this is what we want as a society. Who knows, maybe we do - but not because somebody somewhere is/would/might do the same.
Have some respect for the intelligence of your peers.
There's also cameras scattered throughout the airport which have spinning/swirly LED rings around the lens (which look like this: https://www.hrsid.com/assets/images/art/proven-passenger-que...) -- meant to trick you into involuntarily looking at the cameras. It's flagrantly and unashamedly dystopian.
Honestly, the ‘forces you to look into camera with a straight face’ is awful idea. Many countries already have cams to scan you when talking with immigration officer.
I'm pretty sure those cameras at Gatwick are supposed to track how long the waiting time is, not to "surveil" you in any meaningful sense.
Then they did a really good job of making me feel like I'm in the opening scene of Half Life 2.
I strongly concur that civil aviation needs to be protected from malfeasance. However, all this fancy iris/face-tracking machinery is only useful for, well, tracking and identifying people and doesn't do anything to prevent people from artfully concealing explosives through security checkpoints. Furthermore, it's possible to hide on-body contraband from full-body imaging devices: https://www.usenix.org/system/files/conference/usenixsecurit...
Afaict, there's too much development/use of methodology that has secondary uses for mass surveillance (and isn't even good at detecting explosives), and not enough focus on detecting energetic materials. Large-scale automated iris/face detection on noncooperative human targets is a dangerous tool in the wrong hands and it is frankly disingenuous to develop it (and make its use socially-acceptable) in the name of improving civil aviation security.
Because automated iris-at-a-distance systems that can track and identify non-cooperative human targets scale a lot differently than tetchy border guards. Pervasive CCTV is one thing if it's mostly not watched (and only used to piece together what happened after incidents), pervasive facial/iris-recognition that is watched by computers 24/7 and tracks every single human's movements is a whole other thing. Also frankly, given you recognise the ability of customs/border/immigration personnel to make your day awful, why do you seem fine giving those people even more power and methods to pervasively surveil and track people?
> I'm pretty sure those cameras at Gatwick are supposed to track how long the waiting time is
Even per publicly-available information, this is flagrantly false. Queue management does not require enrolment of the passenger's face/iris data with their boarding pass, as indeed happens at Gatwick. This is explicitly for tracking passengers through every area of the airport:
http://www.professionalsecurity.co.uk/news/press-releases/mf...
http://www.planetbiometrics.com/article-details/i/1850/desc/...
airports that operate common departure lounges[2] must comply with
UK Border Force conditions that are necessary to prevent any
circumvention of UK immigration controls.
This includes a requirement to capture a facial biometric for each
domestic passenger on entry and exit of a departure lounge in order
to verify their identity.
[1] http://www.heathrow.com/departures/security-and-baggage/flyi...[2] AIUI what they mean by "common departure lounge" is a lounge accessible to international and domestic passengers without having to pass fully through to arrivals and re-checking in on another flight, but I could be wrong.
It's thin to the educated few. But for the uneducated masses, they think government is doing this to help them.
The easiest way for a government to get people to accept tyranny is to scare them with a threat ( external or internal ) and then claim the government is taking away rights in order to protect the public.
The power of fear and propaganda is central to government overreach and control. Whether it is nazi germany's fear mongering over the international jewry or the soviet fear mongering over capitalist west or the fear mongering of communist china, north korea, britain, saudi arabia, israel and of course the US, all governments follow the same script.
"If you want to control someone, all you have to do is to make them feel afraid." - coelho
I can't believe the TSA came in _as low_ as 80% a failure rate. The fact that they managed to stop _anything_ done by professionals is sort of amazing in my mind.
> The news of the failure comes two years after ABC News reported that secret teams from the DHS found that the TSA failed 95 percent of the time to stop inspectors from smuggling weapons or explosive materials through screening.
Pretty obvious what they're doing.
Also - how forcefully? Just asking firmly or taking your things off? And what about religious hats, like the Sikh headwraps?
I haven't had them make me do it, but I wouldn't be surprised if you put up a fight.
Also, what do they do with people wearing lots of makeup?
These cameras first went up around the time the H1N1 swine flu was making headlines, and then just never came down. Since everyone is made to present their faces head-on, without glasses/masks, it would be trivial for the government to simply siphon off the data.
having said that, the immigration department recently launched the opt-in self-check-in with facial recognition though
The option is already there for this - TSAPre
2) Requires you to give the government your fingerprints. This is particularly a problem if you have children between the ages of 13 and 18 (old enough to not just get pre-check if the parents have it, not old enough yet to make their own decisions about whether the government should have their fingerprints).
3) Not available at all airports.
4) Not available at all gates at the airports where it _is_ available (even for US airlines).
It's better than nothing as long as item 2 is not a problem for you, but it's a lot worse than sanity would be.
We couldn't lay off the nearly 250,000 active employees of the Security Theater arm of the US Government just because we had an actionable technological solution that reliably worked. We have votes to think about!
I would be interested in the cost-per-life-saved by TSA, and how many orders of magnitude larger than universal healthcare it would be.
1 - www.latimes.com/business/la-fi-aclu-tsa-20170207-story.html
1. Add a bunch of ineffective, but annoying so-called security procedures. Truly annoy people. Take off your shoes and belt!!
2. Offer a way out of that nonsense, in exchange for trading away your privacy.
3. Win.
I fully expect to have to scan my face AND remove my shoes and belt. The TSA doesn't give a fuck, and no one holds them accountable for the delays they've added.
Casinos and even grocery stores have been using facial recognition for years. I highly doubt these systems have not been in place for airports for a long time.
Nothing suspect about this.
Or do you mean you wouldn't even have to scan your passport?
I don't think there's any kind of make-up that would be visible to a camera, but not obvious to the naked eye, and still be able to defeat face recognition.
My plan would be to create a new religion in the United States that makes the wearing of face masks and veils in public an essential tenet. After the traditional new-religion battles over tax status and state recognition have been won in the US, shame the European countries into reforming their enabling laws using the mallet of religious liberty.
The downside is that you would likely have to actually believe--or at least convincingly fake--that showing your true face in public is taboo, and that being forcibly unmasked is a horrifying violation of your privacy and personal dignity.
I recall reading a short story that might be useful as basis, wherein everyone wore a mask at all times, and never spoke without providing musical accompaniment from one of a variety of instruments. The economy was based upon reputation and character rather than money, such that you could get most valuable things just by asking for them, but the shopkeepers had to decide that you were good enough to bear them first. The low-status masks were somewhat standardized and uniform, such that you could be "red goblin" for a day, and only play your pocket xylophone, and thus achieve a measure of public anonymity. Then, the next day, you could be "glorious fiery phoenix" and play your double-stringed lyre with brass resonators, and everyone would know you to be a really special and important person, because the craftspeople would never give away their best work to just anybody. Unfortunately, I have forgotten the title and the author, and web searching is turning up nothing.
And unfortunately, that is a very long-term strategy, which would take many years and a lot of cooperation to pull off. Right now, anything you are likely to do to defeat automated facial recognition is likely to get you additional attention from human cops. And depending on where you are, that "attention" may also include extrajudicial punishment for non-conformity.
http://www.unexploredworlds.com/RealPulp/htm/rpulp145.htm
edit: Huh there's a weird repetition in the middle it looks like, unless there's two almost exactly identical passages.
Since you could only ever get a small group to attempt this it would only draw more scrutiny onto yourself and make your life harder. It wouldn't be enough of an inconvenience for the authorities to care.
The largest group I've heard of is the opt-out of scanners group who refused and always demands xrays. It even made the media. It faded away and now the one or two people that care take an extra few minutes in security getting a pat down. No real change happened.
Isn't that kind of the issue? The majority, some of whom will rarely if ever travel by airline, are fine with this crap because they somehow thinks it makes them "safer" vs terrorists.
>pat-downs
I used to opt-out up until a year ago. I felt like it was the only way I could take some small stand against the security theatre crap. They had just introduced the new 'enhanced' pat-downs - back some guy being much more firm and thorough with his touches and wiping the back of his hand across my dick 4-5 times. That sat with me (still does) and I said fuck it, the system got me, and I do the body scanner now.
Beyond the mass surveillance BS, I'd really like us to take on the state-sanctioned sexual assault that has been going on since 9/11. I've seen grannies and pre-teens patted down too. So ridiculous. Checking the private areas of citizens doesn't make this country safer.
They were doing user testing, and one of the folks they tried it out with happened to be a wiring contractor who was in the building at the time. It didn't recognize his hands; the IR cameras could not see them. Turns out that the lubricant he was using to pull cables through conduit was blinding them.
I don't know what the lubricant was, but hopefully this reduces your search space dramatically :-)
Does anyone know?
> losing your social security or credit card numbers to fraud is nothing compared to losing your biometrics. While you can change those numbers, you can’t easily change your face.
What does this even mean? Last time I checked anyone can take my picture any time I am in public.
Is their theory that someone is going to use a mission impossible face mask to impersonate me and then get me put on a terrorist watch list?
How would having machines doing facial recognition change anything from having humans doing facial recognition?
If anything the machines will be more accurate, so face theft will be less of a problem, but I have never heard of any real life cases of face theft, so I don't actually think it is or will be a problem.
> However, due to the fact that immigrants and people of color are disproportionately represented in criminal and immigration databases, and that face recognition systems are less capable of identifying people of color, women, and young people, the weight of these inaccuracies will fall disproportionately on them.
This is a fascinating finding and definitely a subject of concern.
You can read the research this is based on here: http://openbiometrics.org/publications/klare2012demographics...
What I would point out here is that even if there are race and gender discrepancies in algorithmic performance for facial recognition these systems are being used as a replacement for humans who are not exactly known for being free from racial and gender bias.
At least the facial recognition system is not going to share see through images of you with their colleagues or give you an enhanced pat down because they find you attractive.
Even if the algorithms are less accurate for some groups it should be possible to adjust the algorithms to require a higher level of certainty for those groups so that no group experiences a higher rate of false positives than any other. That is something you can't do with people that you can do with a machine.
> These agencies aren’t just collecting biometrics for their own use; they are also sharing them with other agencies like the FBI and with “private partners” to be used in ways that should concern travelers.
This is be big crux of it for me - that the data gets shared outside of the Government, and I'm honestly a bit iffy on it even within Government agencies. Already the false positive rate for many of the identifying methods used in such investigations is too high; a lot of research is being done as to the validity of most forensic analysis regarding fingerprinting (as well as other traditional forensic evidence) and how the justice system is not doing due diligence in their case research, is not being honest with their findings when presenting to court, and public belief in such myths created by TV dramas has dramatically altered jurors' understanding of such evidence.
Basically, such data getting shared out that makes it that much easier for law enforcement to concoct a case is disconcerting, to say the least, and I'm not even part of the population that would be most heavily affected by it.
If this was just being used in one agency or area of focus (Border Control/TSA), it would be a little less disconcerting. It's not great still, but I suppose it's better than the current passport system for figuring out who you are. (As an expat, my biggest fear is losing my passport and proving who I am in a completely foreign country, much less returning to my own country). To be able to shift my means of identification to something I cannot lose or have stolen from me would be very reassuring, as long as it required multiple live samples to verify.
But the fact that this information leaks out by design into a very dangerous world of Law Enforcement that doesn't care to make sure they've got the right guy is frightening. I've dealt with this indirectly as a certain State in the union decided that I was overdue for multiple vehicle violations, even though I had been out of the country for the last 3 years when the incidents took place; it just so happens I share the same name as someone else who is not a very good driver. It took a lot of talking on the phone and finding a way to fax flight information and passport stamps to convince them to stop bothering my parents with letters and phone calls, since their address is my mailing address. I was lucky enough to be able to tell them to "go fuck themselves" with indisputable proof that I was not in the country; others would not nearly be as lucky, that that's pretty frightening.
Biometrics are basically useless, but they're useless in two separate contexts.
Using them for access control is useless because of this:
> Last time I checked anyone can take my picture any time I am in public.
Lifting someone's fingerprints or face/iris scan is easy, and fooling an automated scanner is not very expensive. Which makes it terrible access control any time gaining access is worth more than e.g. $50.
> If anything the machines will be more accurate
That is definitively false. Face recognition is one of the things humans evolved to do as a prerequisite to survival. We're really, really good at it. The only reason anyone wants to use machines at all is that they're cheaper and scale better. They are not more accurate. At best they may become as accurate, but that is only a headache because it makes their false positives undesirably convincing.
Which brings us to the second context in which biometrics are useless. Trying to match random people off the street to people on some kind of watch list. Because the false positive rate is a few percent, which is several orders of magnitude higher than the ratio of people on the watch list to people in the general population. So the vast majority of people the machine flags will be innocent.
Worse, the people the machine thinks look like someone on the list, it always thinks that. So you implement something like this and it ruins the lives of anyone who looks enough like someone on the list for the machines to consistently flag them.
I am taking the "cheaper and scale better" part into account when I say that machines will be more accurate.
If you want to pick a face out of thousands in a crowd or match a single face to a list of thousands of targets facial recognition is going to outperform humans at the same cost.
> the false positive rate is a few percent
The false positive rate is whatever you configure it to be. The confidence level for flagging someone is a parameter and you can adjust it to whatever false positive rate is acceptable.
> the vast majority of people the machine flags will be innocent
Every single flag will be reviewed by a human who will decide whether or not to take action based on it. This is no different than a human looking at a list of suspects and then watching for matches, except that the computer can use a much larger suspect list, and give the human hints.
> Worse, the people the machine thinks look like someone on the bad list, it always thinks look like them
Every time a person is flagged that must be reviewed by a human. If the human marks the flag as a false positive then the algorithm learns from its mistake.
The idea that people will be repeatedly flagged is FUD unless they look a lot like the target, which is a problem with humans too: http://www.chicagotribune.com/news/local/breaking/ct-identic...
Then you used the wrong word, because cost and throughput are not accuracy. A machine that always returns "true" can be very cheap and very fast.
> The false positive rate is whatever you configure it to be. The confidence level for flagging someone is a parameter and you can adjust it to whatever false positive rate is acceptable.
When that makes the false negative rate ~100%, you might as well do yourself a favor and throw the entire system in the trash.
> Every single flag will be reviewed by a human who will decide whether or not to take action based on it. This is no different than a human looking at a list of suspects and then watching for matches, except that the computer can use a much larger suspect list, and give the human hints.
That's the problem though. Getting arrested and subjected to twelve hours of questioning every time you walk in front of a camera will ruin your life even if you ultimately get released every time. The much larger suspect list only exacerbates things. Even if the false positive rate is no worse, it gets multiplied by a much higher volume.
When the problem is that people make mistakes, making the same mistakes faster only makes the problem bigger.
And of course they'll look a lot like the target. That's what causes the false positive.