German Minister Wants Facial Recognition Software At Airports and Train Stations (www.rte.ie)

An anonymous Slashdot reader quotes a surprising report from Ireland's National Public Service Broadcaster (based on a report in the German newspaper Bild am Sonntag): Germany's Interior Minister wants to introduce facial recognition software at train stations and airports to help identify terror suspects following two Islamist attacks in the country last month... "Then, if a suspect appears and is recognised, it will show up in the system," he told the paper. He said a similar system was already being tested for unattended luggage, which the camera reports after a certain number of minutes. The article reports that other countries are also considering the technology.

Comments

And so it begins.

[Anonymous Coward]

Germany has managed to carry enough post-Nazi guilt to respect privacy, but their underlying culture is fairly authoritarian. While those with a good memory haven't all died off yet, the country does stand a good chance of going from 0 to 100 with a couple large scale attacks.

Which is exactly what the terrorists want, but people are fucking stupid.

Re:And so it begins.

[encad]

West Germany had a lot of this as well,

Just look at the Spiegel-Scandal, Rasterfahndung, especially after the RAF had a couple of hits. Not to the extend of the Stasi, but sometimes not far off. And then there is still the scandals around the Verfassungschutz and the BND.

It's not the first time, that Interior Minister put measures like that on the agenda, but most likely it wont proceed. Some very intelligent people put a lot of thought in Human Rights Conventions and our courts honor them most of the time, beside a lot of pressure build in the press and NGOs. Best example for that is the data retention law, which is already in its second incarnation after the first one got hammered by the Federal Courts, and this one has quite good chances to get hammered, too.

Re: And so it begins.

[Joce640k]

I would say people want safety. They want it so bad that they will believe almost anything if it will make them feel more safe

That's Trump's tactic in a nutshell.

Mexicans, ISIS...etc. He's the only one who can keep you safe. He'll build the walls, he's the only one that can keep them out.

Look at how well he's doing in the polls based on that.

Re:And so it begins.

[dave420]

Because they frequently tell us. That's the entire point of terrorism - to have goals, and to use force or the threat of force to politically coerce people into achieving those goals. You being ignorant of this doesn't change anything.

At least they'll announce it in public...

[guevera]

....these systems are probably already deployed in American train stations and airports, and they won't bother telling us about it. It'll take years just to get details about it through FOIA.

Re:Not a good idea

[WheezyJoe]

Scotland Yard, for one, has a bunch of specialists with a talent for grainy security footage. It's taken a while, but now that super-recognizers are actually looking through all that footage, it looks like the cameras in London are starting to put people in jail.

Where I live, it seems cameras have at least convinced crooks to put on ski-masks before they rob a bank teller or a convenience store. I've got mixed feelings about a world gone all Minority Report, but if you live in a neighborhood where this kind of shit-crime is common, you start to get frustrated at the grainy blob on the 11 o'clock news carjacking a lady at a gas station. It's these assholes who'll make it easy for toothy salesmen to sell politicians on armed security drones, DNA sniffers, cyborg security-dogs, and whatever else crazy shit the future has in store for us.

Create Facebook accounts for all of them

[jfdavis668]

Hack into the system, download the images, and create Facebook accounts for all of them. Facebook is way better at facial recognition. All their friends will help identify them and add even more images.

Excellent Idea, good first step towards ...

[GrokvL]

making Germany great again! With assurances it will only look at terrorist suspects?

Re:Not a good idea

[gweihir]

That is not actually a problem. Because if they actually knew what terror-suspects look like, they could get them by other means. Fact is, this is not about terrorism at all, this is about getting the public used to Big Brother.

These systems do not work and never will

[lars_stefan_axelsson]

Due to the base rate fallacy, systems like this don't work, and never will.

There just aren't enough terrorists to make this worthwhile. Let's assume that one person in a million is a terrorist (probably a high number), and let's furthermore assume that the system get's it right every time it actually sees a terrorist. Let's also assume that it only get it wrong once in a thousand when it sees a non-terrorist, i.e. once in a thousand the system will say "terrorist" when it's not.

With these figures, you will have one thousand false alarms for every one terrorist you catch! I.e. a completely unusable system, that will drown their users in false alarms.

Note that these figures are also completely unrealistically good. Real facial recognition systems that work with willing subjects, are in the high nineties when it comes to true positives, and in the single digit (or low double digit) percent when it comes to false positives. Not 1/1000 that we assumed above.

Now, in relative terms of course a system like this helps. We've increased our certainty from 1 per million to 1 per thousand. That's a thousand fold increase. But in absolute terms it's still unusably bad.

And this is incidentally why we don't screen for most/many diseases in the population. Even with a good test we'll drown in false positives. And the math works the same way for many other situations as well.

No islamist attacks

[Anonymous Coward]

Those were neither islamist attacks nor terrorist attacks, but rampages of two indiduals.
The proposed technique won't help with that.

nope

[Torvac]

a) bild am sonntag is not a newspaper for the 1000000th time, its yellow press
b) in the past there have been tests in germany with systems like this, can be fooled easily without much technical effort
c) that guy is a complete tool, he has no cluen what to do. all he suggests is shit others allready failed with (good thing he hasnt heard about rfid yet)
d) there is no way any government agency in germany could impletement and run a system like this (not even with private companies as partner)

Nothing new here

[bickerdyke]

Face recognition in train stations has been in evaluation ~10 years ago.

And even if we assume that 10 years later they may even be able to find terror suspects in a crowd there, there is still the elephant in the room that they need to be KNOWN terror suspects to begin with!

And whoever touts this as an effective measure against "terrorist attacks" as the ones we had lately, has to completly and willingly ignore that these haven't been carried out by any known "suspects" but some random gullible teenagers have been talked into bringing a knive and stab a few random people. None of these attackers have been connected to islamist/terrorist organisations before their attack, so even working face recognition systems would be useless here.

The German security service tried this years ago

[davecb]

The size of the problem space made it impossible. Any margin of error whatsoever, multiplied by the (number of people you're looking for + the number of people passing through the airport) leads to insane number of false positives. The German Federal Security Service did a trial with Siemens' recognizer many moons back, loved the technology, hoped the number of false positives would be small... and were disappointed. Even with an unreachably high efficiency, it kept tagging grandma as a terrorist.

It's like the birthday paradox: with only one chance in 365 of two people having the same birthday, it turns out that with 23 people in a room, you have a 50% chance of two birthdays matching. A 99% chance if there are 75 people. See http://danteslab-eng.blogspot.... As he notes, if you have a system that is 0.999999 accurate (one in a million), we have a 50% chance of a false positive or false negative as soon as we have scanned 1178 people... meaning for about each 1000 people we either arrest grandma or let Osaman Bin Laden stroll through.

They've probably reported that already, and been told "don't worry about mere mathematics, this is politics" (;-))