In the 80s, an anarchist who wanted, for example, to burn some building, developed his/her plan and at the same time looked to see if there were no listening devices in his/her house. In the late 90s, the same anarchist turned off the phone and used encrypted traffic on the Internet. In the 2020s, we need to reconsider our strategy, intelligence gathering has improved and we must also take this into account.
Originally published by 325.
Note: Enough is Enough is not organizing any of these events, we are publishing this text for people across the US and Europe to be able to see what is going on and for documentation only.
To begin with, let’s see how big data analysis is used. To do this, we need to talk about 3 things: metadata, templates and networks. It seems boring and difficult, but I am not a technician, and I will not bore you with technical language, I will make it as simple as possible.
Metadata: In the context of online activity, “content” means “the message you sent”, and “metadata” means “everything except the content”. So, for example, if you send a text about lunch to your friend, the content may be “Let’s go on lunch”, and the metadata may be “Message sent 01/04/2018 11.32 from 0478239055 to 079726823 using Signal”.
This information is recorded by your phone, even if the application encrypts your actual message. Your metadata is very poorly protected by technology and very poorly protected by law. No matter what country you are in, the majority of your metadata is freely available to special services, regardless of whether you are suspected of something.
Templates: Whether you realize this or not, your metadata has a template. If you have daily work, you can have a very consistent pattern, if there is no such work, your template may be more flexible, but you have a template. If someone wants to know the rhythm of your day, they can do it very easily, because your template is in metadata.
For example: Maybe you use Wi-Fi in your favorite bar on most Sunday nights until midnight, you wake up around 10 am and check your Signal, you use your public transport card to get to class every Monday afternoon, and you spend an average of 1 hour per tumblr twice a day. All this is part of your template.
Network. You have an online network. Your friends on facebook, the people in your phone’s address book, the dropbox you are sharing with your colleagues, everyone who bought online tickets for the same punk band that you attended, people using the same WiFi points as you. Take your networks, combine them with the networks of other people, and the clusters will manifest themselves. Your working community, your family, your activist scene, etc.
If you are in the anarchist community, it is likely to be quite obvious from all of your small network connections, such as visiting one band and knowing the same people as other anarchists. Even if you have never clicked on an anarchist facebook page or didn’t click the go button on the anarchist facebook event, your network is hard to hide.
Now, let’s say you committed a crime, one that would lead to a serious investigation.
Suppose that on Sunday at 3 am, you and your friends go out and burn the house of the Nazis. (Of course, I would never advise any of you to do something like this.) Obviously, the anarchists did, but there are no other clues. You are using a traditional safety culture: you burn records, you try not to communicate your plans near technology and you leave no physical traces.
But since you committed a crime that night, your metadata will be very different from your usual rhythm: you stay in your usual bar until 2 am to wait for your friends, you will not wake up at 10 am and check your Signal or you will Tumblr only in for an hour of the day. You do not go to class. Your metadata template is very different from your regular template. Your friends metadata models are different too. If one of you is clumsy, they can generate a super suspicious metadata signal, for example, the phone turns off at 2.30 at night and is activated at 4 am. You are not the first here.
If I wanted to solve this crime using data analysis, then I would do the following:
• allow a piece of software to analyze the patterns of the local anarchist scene to identify the 300 people most associated with the anarchic scene;
• allow the second piece of software to analyze the metadata samples of these 300 people in recent months and identify the biggest metadata changes on Sunday evening, as well as any very suspicious metadata activity;
• Exclude variations of the pattern with an obvious reason or an obvious alibi (people who are on vacation, people who are in the hospital, people who have lost their job, etc.);
• Would conduct a more in-depth study of those who remained.
That’s right, from the huge number of people I could not listen to at the same time, I can quickly identify a few in order to closely monitor them. So I could find and catch you.
And now what?
If a traditional safety culture will not protect us as before, how do we adapt? Well, I have no answers, but for a start I would say: know your network + know your template.
In the case of the example above: leave the bar at midnight, go back home and put the phone on the bedside table. Check the apps you usually check before bedtime and set the alarm for 10 am. Return to the bar without a phone. Wake up at 10 in the morning and check your Signal. Drag yourself to class or ask a friend to travel with your travel card and do not use technology in your home while a friend travels with your travel card to class. Stick to your template. Never turn off the phone.
You can also manipulate your network, but it is much more difficult to do. Do not use the smartphone in general and abandon all social activity on the Internet – this requires serious motivation. Knowing your data template and making sure that it looks ordinary is much easier.
Some of the old rules will still apply: do not talk about crime around devices with microphones, do not brag after successful actions, etc. Other rules, such as “turning off the phone when planning illegal actions”, need to be changed because their metadata looks too unusual. No one else disconnects your phone. We look suspicious when we do this.
This is just one idea on how we could update our safety culture. Perhaps there are other people with different, better ideas about updating our safety culture. If we start a conversation, we can get somewhere.
Finally: we need to continue to adapt
As technology changes, more information emerges, including data that we have very little control over. Smart-TV and advertising in public places that listen to every word that we speak, and the tone of our voice when we speak, these are examples. Currently, data analysis projects use license plate reading software to compare vehicle traffic patterns. It says a lot that they may soon be ready to do the same with facial recognition, after which the presence of our face in the public space will become part of our metadata. Additional information means more accurate data analysis. Our metadata may soon be too extensive, which is too difficult to fully reflect and mirror. This means that we will need to adapt our counter measures if we want to hide something.
How do we keep all this under the radar? I don’t know. But let’s try to understand this shit. These are some first thoughts on how a safety culture should look like in an era of modern analysis of large data sets, and I would be very happy to receive additions from comrades who have thoughts on this.
Also, feel free to distribute and redo this text without references.
Support Enough is Enough! Donate!
Donation for our work in the Enough is Enough info-shop and our independent reporting on our blog and social media channels.