My nightmare is coming true


There’s a lot of heat those day in France about a call for project made by the RATP (the more or less public organism which is in charge of managing all the private and public parts of the Parisian and Île de France commutation system). It’s a simple project that would be used to help regular customers to automatically pay their monthly fee using a facial recognition system.

If you’re not Parisian, or a suburban, you have to know that there’s actually two legal way to use the commutation system. One trip tickets, or monthly pass. Those monthly pass are either paid directly by taking the money from your bank account, or every month by the mean you choose (cash, check, credit card). Those pass are either anonymous (you buy them 5€ and you write yourself the data on it, so they don’t have your name in a database) or linked to a account in a database (but the pass itself is free, and they geenrally makes you a discount). This whole system is the Navigo system and is, surprisingly, quite safe (given the security of other system you may have met).

So, now they want people in their databases (those who, most probably, have an automatic monthly paiement already in place) to automagically pay their pass using a facial recognition system (so, instead of the credit card they’re not currently using). It must be this because, you know, other people don’t want to have an automagic paying system (or they would already use it).

You’re wrong!

If you think it’s only about payement, then you’re wrong. The people who can benefit of a real gain of time (we’re speaking about half an hour once a month), have the year pass and they do not really care.

So, now that we agree on the fact that this whole "it will be easier to pay" is a lie, let’s look at other things.

First, when people found the document (online, in a public dropbox, no encryption – seriously people???) and made some publicity about it, the RATP did remove the call for project from their website and, a little bit later, the project. But internet is cool, and the streisand effect still works fine, so you’ll find it everywhere. It juste mean they’re ashamed of it or do not want people to be aware of this project.

Another interesting point is to look at all those European Funded Project into which RATP – as well as arms manufacturer – is involved. The most interesting one I found lately is the VANAHEIM project, developped with the RATP equivalent of Torino. They have now a fully automatic system which can do machine learning and can interpret unusual behaviour (such as someone falling in stairs, someone jumping the fence, a lost tourist or – and it’s used as an example – someone distributing some flyers to people) and launch an alert.

These kind of systems have been widely used in the RATP subway system for a while, and they’re ineffective because, in case of agression, the mother fucking cops arrive in more than fifteen minutes.

You’re wrong!!

If you think it’s a privacy invasion. It’s not about privacy. You’re privacy is at stakes in the EU Parliament, and you should speak about it with or, but this is not the issue at stakes with this kind of project.

I mean, there’s already way to much CCTV cams in the subway (just getting inside and outside of the subway system makes me going through 30 of them, for a 100m trip that’s a lot).

Besides, since you’ve bought the fame, you’ve got a tracking device in your pocket making you leaking a lot of data about your errands, who you’re with, and the call you’ve made (just search for FinFisher ‑ and don’t think they’ve stopped doing that because it has been exposed).

There’s also the fact that you want to check on your (girl|boy)friend, where your kids are or if the elders one are ok. You asked for it. You beg for it. You want to spy people around you instead of trusting them, and you can now buy a GPS tracking devices to tap your partner for less than 100€. And after that, you complain about someone else than you doing the same thing. You deserved it.

But in this case it’s not about privacy, the RATP can barely do more harm to your privacy than what they’re doing right now. They’re currently linking biometrics data (picture), names, travel history for the last 48h (that’s what they claim), bank account and now, since they’re selling a device to load your pass from your computer, they even got the address and details about the computer you’re using at home (and with the possibility for them to install a software on your computer). So no, this facial recognition system (one another) is not about your personal data or your privacy.

It’s about being able to analyse your behavior, the way you dress, the way you walk, and to automatically launch an adequate response (I’ll go for cops right now, but it won’t last long before they send drones or before an automatic fee being paid directly by your bank without your knowledge).

You’re wrong!!1

If you think it’s only a RATP issue. I mean, Thales is working on it. They are weapon manufacturers (and sellers), they build missiles and drones. The transportation system is just a sand box and an experiment for them. The police will use it soon (what, you’d expect they’ll leave human behind screens to interpret CCTV data? Well, with all the CCTV – heck, sensors there’s microphones now and I expect infrared camera that interpret feelings or electromagnteic scanners who scan backpacks, like there’s already some of them in airports customs system).

If you’re interested, there’s a whole bunch of similar project funded by the EU, just look for INDECT.

We know that a lot of private companies are already spying on everyone, and they’re not even limiting their job to borders (hint: borders don’t exists but to jail you), look at what Gamma is doing, or Qosmos (with HackerTeam – oops, another Franco-Italian alliance in tech) or even Google or Facebook (hey, they’re also doing facial recognition).

Those systems are used worldwide to detect insurrectional comportment, to identifies bad citizens, ask China or Belarus about that. And now they’re going one step further. They’re just trying to remove the human from the decisional system because human is expensive and may have one day a moral issue with doing this job. (hey, go look at the spyfiles if you do not think they’re doing this)

We now have systems which can analyse and track "bad" citizens in the world on one hand, and, on the other hand, we have assassination systems (drones) ran by private corporation (asks the CIA who’s operating their drone – for the record CIA is NOT a military organistaion it’s just a civil one), to kill people without trial wherever in the world. They’re working on automatic drones (because you can’t rely on the stability of communication in times of war) which will be unstoppable once launched, just too ensure the result of the mission, whatever electronic warfare is deployed on the other side.

I’m just wondering at what point they will connect the two systems (and then the Santa Claus will be happy, all the "bad" kids will be killed once detected by the system). I mean, they will, I just don’t know when it will be done. We do have "analysts" who are now filling data gathered by autonomous system to determine who to kills first and who then launch an automatic assassination system without supervision, or accountability. The day where the analysts will disappear isn’t that far.


If you think it is a nation-state related issue. The nation don’t need an imprecise and automatic system to kill people without trial. If it’s a democratic state it should not do that, if it is not, dictatorship are already effective at that (the history is full of people mass-murdering their population without any computer aided system).

Who’s gonna benefit from that? Come on, it’s easy. It starts with a B and end with a USINESS. Business. Money. Profit. Control. This is what corporations feeds on.
Arms dealers need war. Security dealers need insecurity. Cyber-defense sellers need cyber-threats.

This is the rise of new automated systems. Google has already built the root of an AI and they paved the way to mega-scaled computing to exploit and make sense of data to answer the best way possible to questions (such as "what is Finfisher?"), Thales (and Safran / Morpho, and a lot of other companies) now have an efficient system of analysing the comportment on a city scale. We do have autonomous weapons. The financial systems doesn’t depends on human intervention (another question for Google: "What is High Frequency Electronic Trading?"). The legal systems of all the countries are threatened by the operators of those not yet sapient AI (continue asking Google some questions: "What is ACTA?", "What is TAFTA?") because they just want more control.

I said efficient? Well, it’s not. There’s a lot of false positive. But those companies don’t care, all they want is the system to be efficient enough to be sold. The fact that thousands of people have been killed by drones doesn’t mean that those people were all targets. They just were at the wrong place at the wrong time. Or the operator was high or drunk and just thought he was playing Medal of Honor. This is probably the worst part.

This is a new era. The rise of the machines. And, according to James Cameron, it won’t end well. I’m not opposed to computers and technology. I’m opposed to technology who take control of my life and yours (because then they can control my life)


Thanks to the whole karmeliet team who did a lot of correction here