JetBlue explains to a passenger how it got a photo of her face
boingboing.net
boingboing.net
It is of course legitimate for a democracy to check things like which plants are coming across the border and to deny access or apprehend war criminals and the like. However, the USG does not have a current record of using its powers for good recently. Example: the muslim ban. Anyone who managed to get around that is a hero even if it was illegal because it's an immoral policy.
Tighter control means tighter control of dissent. Want to find that soldier with those leaked documents? Want to locate that journalist with news of unseemly interventions in central America? Want to find those people of the wrong race or religion that are the convenient scapegoat of the moment? Sure, let's make border control more precise. Most people are actually decent and should have default global freedom to move and work by birthright. US border control is about denying people entry, but the main effect is denying ordinary people with a minor effect of deterring crime.
Hold politicians accountable for what they are doing, from the local community to the President. Defend the freedom of the press, especially when Trump goes on a rampage again. Vote for someone better than him, vote for someone actually progressive. Convince family, friends and co-workers. Organize yourself and head on to the streets.
For those willing to practice with arms, join Redneck Revolt. Protest against Nazis taking the streets, protest them wherever they appear and want to spread their propaganda.
Oh, and a look towards France and the Yellow Vests is never wrong. The willingness of the French to go on mass strikes and the occasional riot has proven to be very effective against the worst of neoliberal politics over the last decades.
We criticize them just a little, by principle, but as soon as the tech proves effective at controlling population, we adopt the tech "because it works".
"But at what?", one may ask...
If you include economic causes, you might need to include the excess mortality of systematically impoverished communities in western nations — and I don’t just mean “PoC in the USA”, as I can point to a UK study claiming the government’s austerity policies since 2010 have caused 120,000 excess deaths: https://www.bmj.com/company/newsroom/health-and-social-care-...
Also, if those deaths were related to economic policies, you probably have to weigh them by population size.
Obviously that’s still a million executions more than most; I am not defending him, this is just about making it an apples-to-apples comparison.
Numbers from
https://en.wikipedia.org/wiki/Excess_mortality_in_the_Soviet...
I'd say if famines are the cheapest way, many dictators with a genocide mind set would do that if possible.
We have learned a lot about building stable power structures in the last 50 years. You don't need to have a death grip on everything in the country. A much softer touch is the technique of choice of modern oppressive regimes.
Why disallow other parties when you can just make sure the right party always wins?
Why get rid of capitalism if you can just pick the winners?
Why censor information when you can drown the airwaves in fake news?
Why purge dissidents if they have no power to change things?
Why control what people do when you can control what people want to do?
> It took her photo, comparing her picture to a preloaded photo databased of all the passengers with tickets on this flight, and then she got a check mark, indicating she was cleared to board. The whole process took about five to six seconds.
> The Aruba experiment is expected to last somewhere between 45 and 90 days.
> "We're basically capturing that picture at the boarding gate, providing it to U.S. Customs and Border Protection. They're identifying the traveler," Farrell said. "It's actually the U.S. government that's implementing the biometric matching system that does all the hard analysis and crunching of the data."
> A biometric exit system to track non-U.S. citizens using their faces or fingerprints has long been a congressional mandate, particularly after 9/11, to improve border security and identify people who've overstayed their visas.
> Crockford says people don't know how CBP might share that data with local police. CBP insists it'll discard photos taken of citizens, and only keep a database of non-citizens.
> Customs and Border Protection has been piloting similar biometric tests at airports in Atlanta, New York and the D.C. area. And, according to a customs official, the goal is to deploy facial recognition tech widely by early next year.
I guess it must be comparing with their passport photo? (Or visa photo, for foreigners?) It doesn't seem much different from the way you're already required to show your physical passport and they check your photo manually. It's also always been pretty normal to have your photo taken by countries when passing through passport control. So I don't really see what the problem would be here?
[1] https://www.wbur.org/bostonomix/2017/06/21/jetblue-facial-re...
That said, even as a libertarian, privacy-conscious person, I'm not sure I have a problem with exit controls like this. It's important to know who is crossing the border, in both directions.
But why do they even need the photo? Presumably the name is already on the ticket... surely they can detect overstays from names + address alone, not just photos?
Why?
And if there exists an important reason to know that someone is exiting the country does that same criterion not also apply to exiting a state or city or neighborhood?
I am not being sarcastic; I am genuinely unable to come up with a reason it's important to track who leaves almost any place, except in highly unusual cases (e.g. "OK guys, chatmasta has gone; let's get to work planning the surprise party for when he gets back")
The airlines are giving out all the information, that's why there is no need to do an exit passport control.
From the US? I can't think of any outgoing passport control at all; not by land in a private vehicle, nor by boat.
So France has basically no idea who shows up from Germany or leaves via Belgium. The countries on the external borders are supposed to handle all that.
Now I suppose it's possible there is a lot of data exchanged in real time and so on, but it's also possible that is a complete mess and nobody really knows anything, especially about citizens (GDPR).
I have not crossed that many non-Schengen land borders, but as far as I remember it was the same every time, when going from France to England through the channel, Belgium to England by train, France to Switzerland (before Switzerland joined Schengen) and Canada to the US.
I have been to a few airports where you explicitly clear US customs at the departing airport before leaving for the US — Vancouver, BC and Dublin Shannon come to mind.
If you've overstayed your automatic 90-day Schengen area travel visa, a common hack for backpackers and such is to intentionally leave the Schengen area from someplace like France that's generally less strict about counting exact days than many other EU countries.
To be fair, I don't have much experience flying from the Schengen area to non-Schengen area non-US countries. It's totally possible this is all rigamarole that the US government pressures other countries to do.
I have heard of friends of friends who overstayed a visa and get found out on the exit gate (usually bringing a lawyer/embassy official to talk through things). This can lead to a (temporary?) travel ban to Germany and a fine.
And those maximum visit lengths are counted on an annual basis. So you can only visit the US 179 days out of 365.
So they need to know when you came in and when you left. It's not good enough to just leave and not tell anyone.
But institutions like governments only “see” via data collection, and in general are happy to collect whatever is technically and socially feasible, and people involved in setting up databases seldom stop to think about the risks or trade-offs involved in collecting various types of data, or whether that data is meaningful or useful. https://www.ribbonfarm.com/2010/07/26/a-big-little-idea-call...
Staying for more than six months is usually one of the main measures to determine if one is resident and should pay taxes in the country, so there is at least a very good reason (although I don't know how much this applies to the USA specifically).
Finding people who rob banks at some point and punishing them is of great interest to people fearful of robbers who like disrupting robbers’ lives for no practical purpose because they gain sadistic satisfaction from cruelly exercising power over other people.
You punish a robber ostensibly to stop them from robbing again, lest you end up the victim of a robbery. What does being the victim of someone's unlawful presence even mean?
There is nothing unique about the US in that regard. Most countries have a limit on how long you can visit.
This isn't always a good thing, of course, as there's less due process involved in handling such situations. Unlawful entry is a misdemeanor, lying to a border guard is a crime, and unlawful re-entry after deportation can be, also.
[1] https://www.politifact.com/florida/statements/2017/mar/15/fl...
I think your "who is hurt" approach is weak though.
If a country can't enforce its own laws - or worse selectively enforces them - no one is obviously "hurt" but it erodes trust in the systems and process itself. It leaves everything up to the judgement of the people on the ground and damages the concept of equality under the law. It's a dangerous game to play.
I was suggesting one has a clear victim and the other doesn't so it may not be directly comparable. The "who's hurt" question was something I'm curious the consensus answer to rather than an excuse for not enforcing the law.
Not saying one way or another, but the argument is that ICE was only formed 16 years ago and that the U.S. was not 'overrun' before then. The Border Patrol etc. would still remain, just the Stasi-like tactics of removing people from their homes in the middle of the night and holding people in cages would disappear.
At least that's how I've heard the argument.
Saying "only formed 16 years ago" is only right among the "well, actually" crowd.
The difference in 'effectiveness' is sort of like the difference between the reach of the LAPD intelligence arm vs the CIA.
I believe the US exit controls are currently based on airline passenger manifests (not entry databases of other countries).
Also de facto limitations in highly violent places like Syria, parts of Iraq, Somalia et al, but I assume you were referring to a de jure right
With this development of technology, it becomes possible to even check who is crossing your neighborhood, and you may lose your freedom inside your own country, if a government decided to go in that direction
Obviously a great country to emulate.
This happened to me leaving London and it was a lot of fuss over nothing. I (American) had given my forwarding travel plans to the border agent upon arrival and I didn't notice but she marked my passport with a visa for an exact stay (just the weekend). Except she mistakenly wrote my exit date a day early. It took quite a bit of explaining to convince them not to put a giant "DEPORTED" stamp in my passport and even though they eventually believed me that it was a clerical mistake they definitely kept track in their system because I was questioned about it several months later when I went to the Falkland Islands.
Anyway, I think facial recognition makes sense from a security perspective, as an additional layer of defense. You can make assumptions about what faces you expect to see from the passports that are scheduled to fly. If you see an unknown face in the boarding area, it's reasonable to flag that person for further review.
*US citizens flying within the states.
- Hannibal Buress
Casual article on it: https://www.king5.com/article/news/local/how-to-get-on-a-pla...
ADDED: It appears that as of Oct 1 2020, this sort of exception will no longer be possible.
It doesn't explicitly say "no exceptions" but the plain text is certainly that you will need appropriate ID (Real ID/passport/etc.) to go through TSA when Real ID requirement fully kicks in.
The inspection is not a big deal and they are prepared for this situation. It’s a shame that the laws that control this are classified (cf Gilmore v Gonzalez) so we can’t know what the actual rules are.
It’s a “solution” in search of a problem.
According to quick search there’s no specific resolution requirement for the image in biometric passport, but size should be min 12kB to have recognizable image.
I noticed that these gates were quite slow ID'ing me the first couple of times I went through them but now I hardly have the time to see my face flash on the screen before the gates open. Maybe there's a human check on initial passes. Either that or their code has gotten way better or they store data that helps matching after a few uses.
I'm also flying to Kenya and other E. African countries so it wouldn't be surprising if they flagged me.
My guess is the app probably just approved anyone or nearly anyone. It's unlikely someone without a boarding pass would try to get on and it's likely that anyone with a boarding pass who got disapproved would be pissed.
Or who knows, maybe homeland security did a mechanical Turk and had someone at the other end looking at faces and checking off names.
No not always, only recently.
> I don't really see what the problem would be here?
You’re kidding right? There’s multiple privacy problems. Sounds like you either benefit from mass surveillance or have been conditioned to it.
No, not kidding.
There's a legitimate government interest in identifying people entering and leaving a country. This isn't Facebook, or even cameras on street corners -- this is a government responsible for enforcing legitimate immigration policies at a legitimate location (the border) for a wide variety of reasons (safety, economic, etc.), which absolutely requires knowing who is who.
I don't know any serious thinkers who claim otherwise.
And I don't see a particularly big difference between getting my photo taken for official ID's every few years (passport, driver's license, visas, etc.) and having it taken when I need to prove my identity with one of those.
The government already has photos of me. They already know my flight. How could it possibly affect my privacy that they also have a photo of me when I get on the plane?
I'm just not seeing it.
(Unless you're opposed to the whole concept of passports and photo ID's in general, but that would be a pretty radical position.)
And I'd wager the reason they're doing this is because while a human can make a decent determination of whether the photo matches the person in front of them, computers need a larger sample to get the error rate low enough for broad use. That's great for CBP and airlines that can employ fewer personnel, and great for travelers who can breeze throughª boarding and customs. But imagine how it will go when you are one of those 0.1% errors and the system refuses to let you in.
ªassuming companies/govts decide to increase throughput instead of decreasing cost, or that it's such an improvement that it can do both.
Wasting taxpayer money in photo storage is a laugh. That isn’t the same category of argument as the idea that is is “overreach.”
It would be entirely possible to have a benevolent dictator, he might think he is doing everything for the benefit of the people - and he might even be right. This situation is fine, the issue is what happens when that dictator dies.
Similarly with mass surveillance, if used with restraint and for the right reasons it could be completely fine. The problem is what happens if the situation changes, there is now a massively powerful tool available with little recourse for change.
I think the issue is not the situation as it is _now_, it's the risk potential it inherently has that is terrifying.
You lost me at this straw man. You are tossing around nebulous and vague concerns. Would it be fair to say your concern is potential abuse of your photos by a government entity?
It's fair to say my concern is potential abuse of any system in a fluctuating context.
It's not just the picture, it's the databases and the veracity of mass data base mining that can come from it.
To better explain:
Right now powerful entities have files on me. My actives. And yours. The collection, use and sharing of these files between these powerful entities violate our rights, even if only in intent (though arguably more)
Right now these files are mostly loosely created, probabilistic based associations.
If you've ever used any ID verification systems, you'd understand this (think lexisnexis)
So, they sort of know, but don't fully. Good. Even if these entities have shown the desire to violate my rights, they can't do it well because of issues related to data collection. I like that.
So the way I see these advancements is... great, now these entities are getting mass fed pictures that are 99.999% a correct association. This could be used for facial recognition software tie ins to say, CCTV systems. As stated, you need more than one picture to increase accuracy.
I don't want them to have it.
Right now, their interest (in violating my rights for profit and power) and my interest (in staking claim to the rights my forebearers fought for) are oposite.
However your concern about building a face database is valid. Note that China somehow can do it with only your Wechat profile. Maybe they are better at ML than the US. They are already automatically fining jaywalkers by face ID. And I assume worse.
Easy to say. How about evidence. I offer all the evidence from Snowden and other whistle blowers to beg the contrary.
To me, it's very obvious unless you are intentionally not wanting to see it.
Notice I said if only in intent, I'm not here to split hairs on legalese. Intent in things like the Fourth Amendment to me is very clear.
I live in an area where when I go for weekend drives, I have to stop at border patrol check points where I am forced to stop and answer officer's questions. I'm not even traveling between states and I'm over 50 miles from the boarder in those points.
Trust me, I understand the legalese they use to justify it. They define a border zone that I think goes a hundred or two miles in land, which means it covers something like 70% of major American metro areas and allows the gov to basically conduct illegal searches by creating a loop hole.
But that is the point. My rights are routinely violated with impunity. People who have stood up to this get literally crushed. Clear whistle blowers are declared spies. It's them working to undermine my rights. Rights that people literally died so I could have. I think it's wrong to let those rights go. It's why I support things like the Electronic Frontier Foundation.
This is nothing new, only new for people that don’t live near a physical border in the US.
The government clearly has missed a few people coming in and out on Visas. What is a way better than the current border protection method that also protects everyone’s privacy?
Would your argument against street corner recognition systems have been different before and after facial recognition was required at the border?
You mention that having your photo taken at passport control is “pretty normal”. Do you remember when it was not? Would you be more concerned if we had gone straight from a situation where no pictures were taken, to rolling out a facial recognition system at the border?
Government does not already have photos of all individuals who cross street corners (there is no legal requirement to have ID) and government does not record a list of names of each person who crosses the street. So, cameras at street corners have a much larger delta - between no photos and photos and between no internal location records and having internal location records.
This reeks to me of another situation where we're trying to put the cat back in the bag, and that has basically never worked (except maybe nuclear disarmament, and even then... barely, and largely because the capital costs are so extreme).
Surveillance tech is cheap and widely available and will only continue to move in that direction. Attempting to make sure no one has a photo of you without your consent seems hilariously impossible to me. Focus on the outcomes instead.
For example road side license plate readers sound cheap, but unless a cop is next to it it’s really hard to catch the car that tripped one.
But the London cops have been caught falsely accusing, arresting and holding people for weeks based on facial recognition (based on both wrongly made identifications, and wrong input pictures).
People have had their own medical files used against them in court by youth services and mental health services departments of their own governments.
People have been convicted based on factually incorrect information entered into police databases, then corrected in one place, but not the other.
The government keeps getting caught using privacy laws in ways not intended: for instance, to hide the identity of cops that beat up someone for racist reasons, preventing them from getting sued.
IRS has access to almost all databases. From medical files to bank statements. What they use it for ...
We all know WHO needs to be prevented from merging databases to increase people's safety. And it's not Facebook, it's the police, health and government departments. We need a right to delete any information about yourself AND ALL LINKED AND COPIED VERSIONS in all government database, and all information sharing needs to be "default deny". No government department has any business whatsoever in your medical files. All information provided by the government in court needs to be subject to invalidation by counter-experts. And it should be default practice to have all government information and copies of it that involves any person directly or indirectly "time-out". And to prevent government abuse of privacy laws, anyone working for government needs to be identified on request in any record that involves that person acting in any official capacity.
Some people don't mind these kinds of surveillance. You might staunchly disagree but you have to accept that it's a valid position.
What if I don't accept it?
Is it valid?
At some point we have to determine what is a minimum acceptable level of cohesiveness in though to take someone seriously.
I'm not saying anti-privacy people are in such extreme, but I don't like the 'everything a valid position in the post-modern relativist world'. The line has to be drawn somewhere, and some people might draw it further in or further out. But we all do it.
I don't know of anyone who likes sticking needles in themselves (aside from addicts, but that's another story), only those who understand the value of what's in them (I suppose that addicts would fall into that category, too.)
I distinctly remember my mom explaining why I was about to get stabbed with needles when I was four. I didn't mind.
Well, now they can check your identity anywhere with a camera removing all agency. Yay!
So only non-citizens would be able to "use" the face-as-boarding-pass "feature"?
There is no one clear process anymore.
Last week I went to board a flight and checked in using the phone app. At the airport I needed to get a baggage ticket so I used the machine there for that, but it basically had me go through the whole check-in all over again. On top of that the first machine errored during the step to pay the baggage fee so I ended up doing that again with another machine.
Afterwards I went to drop off my bag and was told the tag was not "active" and I'd have to go to the desk.
What the hell is the point of any of these electronic options if at the end of it all I still have to wait in line for an agent, stand there while they "mmm-hmmm" at their computer and avoid eye-contact with me for 5 minutes just to finally scan the tag I had printed and attached myself?
Checking the list via facial recognition adds extra dimensions to the problem, but is the same core problem.
Shh! Don't let the terrorists know!
So you should check your flight and gate number for your own benefit.
Not to mention the chance would be very low, you have to go the wrong flight and the system have to have false positive at you.
When dealing with large systems, even small probabilities regularly occur.
And here is a prediction, getting a record of faces and fingerprints, and tracking movement is on the agenda right now. In the future there will be a equally focused agenda to acquire records of peoples DNA. In countries where there is nationalized health service, it will eventually be required. Aquisition will begin when the cost of storing, acquiring, and searching that DNA information becomes cheap enough.
https://en.wikipedia.org/wiki/Biometric_passport
https://play.google.com/store/apps/details?id=nl.innovalor.n...
In Australia, you don't need an ID for a domestic flight; and non-flyers are permitted to walk to the gate, no boarding pass is required to go through security.
Not sure if my passport chip is broken or the system isn't full proof?
> Is DHS trying to build a national database with all of our information?
> No. REAL ID is a national set of standards, not a national identification card. REAL ID does not create a federal database of driver license information. Each jurisdiction continues to issue its own unique license, maintains its own records, and controls who gets access to those records and under what circumstances. The purpose of REAL ID is to make our identity documents more consistent and secure.
Actually, it was probably done in a fraction of a second.
It implies that the devices that JetBlue are using for boarding, already have all the photos in them. If this hardware belongs to JetBlue (rather than CBP/DHS/etc), then one would be correct in saying that JetBlue has the passengers photos. So we really don't know either way how this works.
I'll also add my personal anecdote: Recently I flew internationally out of SFO (I live in the Bay Area) and CBP was there to take "exit photos" of everyone with similar style devices. This was not part of the boarding pass process as we still had to do scan tickets afterwards.
> "We're basically capturing that picture at the boarding gate, providing it to U.S. Customs and Border Protection. They're identifying the traveler," Farrell said. "It's actually the U.S. government that's implementing the biometric matching system that does all the hard analysis and crunching of the data."
So I'd say it's more likely that it is CBP who has the database (and not JetBlue).
on the other hand, for cryptographic hashing we want very similar but inequal inputs to be hashed to wildly different outputs, so a useful hash function should be highly nonlinear / discontinuous, small changes in the input leading to huge changes in the output.
if we're hashing in machine learning, we probably want the hash function to be (crudely) continuous functions -- with "similar" inputs hashed to similar values, so we can operate on the hashed values to cluster or classify things. we dont want a bit of "noise" in the input to produce a massive difference in the lower dimensional feature space.
If there is such a person, you could sign up with their name / birthdate and walk through the control points without having to show your ID (which you didn't bring "because you thought the face ID was sufficient").
https://twitter.com/RealPressSecBot/status/11232845870635130...
Also, the government organization is not foolish enough to build such a service they get to be blamed for. They gave/sold that data away.
Just the face is of no use, the identifying information that goes along with it is also given. So at this point any company in the US can just query your entire info from your photo. Privacy is dead and it is provided as a service by the government. Wonderful.
Question is can anyone get this data?
So you're saying, they're lying. Which might be true. But you'd need to prove that.
I've read repeatedly on HN over the past couple of months (re: Boeing and Facebook) that these corporations are required to lie about the naughty things they do. The shareholders' interest must be protected at all cost.
The corporate bogeyman is a great story but I think you’re being a little unrealistic here. These systems typically allow the picture to be sent in with a name, and the system returns a success or fail response.
Please, offer me alternative view points rather than calling my theory FUD.
It’s true that we have no guarantee that the airlines won’t save our photos along with our ID information, but they _could_ already have been doing that for years. There just doesn’t seem to be that much reason why they would.
It seems more likely that the CBP system compares against passenger manifests of upcoming flights. Which isn’t necessarily more intrusive, but it is a bit less straightforward than a simple yes/no query.
It's an airline operating at an airport. That operation is already inseparably interwoven with number of other organizations - including the government - just to move a tube full of people from point A to B.
>much so a government organization that is not known for building highly scalable services
In that case, it's good for them that they have such a highly predictable load target. Airports have a fixed number of gates and a maximum number of planes ATC can smoothly get into the air in an hour. There's your upper bound. Flight plans and ticket sales set the normal band and refine the performance target. Load can be well understood, and they have lots of visibility into future demand changes.
>This is a crucial part of their process
It's a pilot program. As much as I doubt they'd let this system ever be a single point of failure under normal operation, an experimental system certainly has a fallback plan for boarding.
If the government servers go down, JetBlue can fall back to checking paper boarding passes.
At this rate, in 10 years we will see even more "conveniences" like for example automatic face id in shops and on websites that pull your payment information and can authorize payments from your card/account - with no setup.
People will say "Kinda spooky how they can just take my money automatically but it's convenient" and you'll be there to say "It's kind of surprising that we're still surprised by this." ;-) I hope!
That will be underpinned by even larger surveillance and data collections on people, that can do a lot of harm when/if data is leaked or the systems are breached.
https://medium.com/matter/how-i-requested-my-photographs-fro...
It seems they are simply asking for identify verification by sending the photo to a remote DHS server for verification. How is this different than showing a ticket agent your passport and having them scan it? That’s the same photo in question. GDPR doesn’t get you out of having to verify your identity in person against a government photo for boarding a flight.
In fact, one could argue the method in question is actually more secure and private because you never actually have to share your passport photo directly with JetBlue.
But surely, the government has plenty of cameras in customs areas and face recognition has been a thing for a decade.
Maybe the TSA does it, of course.
AFAIK it should apply to entities that share personal data with third parties, for instance.
They got the images from CBP.
The airline isn't taking or storing photos of passengers who don't opt into this process, but it is presumably sending the government a list of all passengers prior to boarding so that photos of passengers who opt in can be verified. Would this violate the GDPR?
I absolutely do not believe the tweets from a company PR representative as to how their technology actually works.
They could also build the same database 10 years ago by taking your picture as you scanned your boarding pass.
I can't say I'm certain it is fail-proof, and it defiantly can feel spooky but not everybody is lying to you all the time.
From the article the PR quoted, I arrived to the relevant executive order that facilitated this, section 8: https://www.whitehouse.gov/presidential-actions/executive-or...
But you're right, and the process is surprisingly reasonable as long as the implementation is safe