Well, that’s all well and good, but what did you do in the 2010s, other than have a kid? Well, I requested my files and created a website that was about access and transparency regarding how the Canadian State (i.e. The Crown) monitors its citizens (who are actually subjects) and I discovered a lot of bullshit surveillance. The problem was that since this wasn’t news to anyone, and the fact that when you talk about the actual bad things Canada does, people point to south of the border and say “The US is worse” like that somehow magically absolves the colonial nation-state of Canada some guilt, you have to wonder, WTF was the point?
So, I eventually took down paroxysms.ca, because it had a stupid name that I took from a Wikileaks Cable Dump, and I decided to just be a parent for a while, because special needs parenting is hard. There were some other serious blows that happened as well, like certain people turning out to be completely abusive shits and dissolving all good will and energy people many people had in this space.
BUT I still showed up and spoke out against the C-51 and it’s successor. I still told the Parliamentary Committee against Surveillance that the use of Stingrays near prisons is bullshit and that not only prisoners but their neighbours would be victims of this surveillance. But it’s much harder to be a critic of the state when there’s a Liberal government run by Trudeau than when everyone views Stephen Harper as the Big Bad, even though absolutely nothing at all has changed.
So….fast forward to today. Instead of worrying about Stingrays, Telco collusion with the government to spy on absolutely everyone and the fact that if you have a phone you have zero privacy, we’re going to talk about something far worse, which is the use of Machine Learning and Facial Recognition by the Government and their intelligence agencies and how that is being used to conduct surveillance on everyone en-masse and to wage war against the Indigenous peoples of Canada, the environmental movement, immigrant communities deemed to be a threat any anyone who is considered a Critical Threat to the State.
So, yeah, some pretty on-brand stuff for me.
Now, when I first got into the ATIP game, it was BRUTAL. You had to mail in forms by hand and you received nothing but hard copy documents with redactions that you had to scan by hand. Nobody knew what anyone else was ATIPping and many journalists liked it this way because it meant that they could scoop other journalists by asking the right questions. It’s also STILL brutal because there’s tons of laws where the government can go and tell you to pound sand and use the laws to exclude certain items from the ATIP documents. Furthermore, documents were prepared for the person and were not prepared for general consumption. For example, when I did a request on the Olympics, my name would be the only one not redacted from a list of Anarchists under surveillance in Vancouver. I would have to ask for clarification as to how many people were on the list and this would be another $5 just to get a number. I would also make informal requests to try and get the original JPEG photo that was used to take my picture to reveal the model of camera that the Vancouver 2010 Integrated Security Unit – Joint Intelligence Group (V2010ISU-JIG) used to take my picture.
Now, a lot of it is still brutal and sucks, BUT there’s a search form so you can see what other people typed in:
Now, when I looked up “Machine Learning”, there were a lot of 0 pages turned up for CSIS, where CSIS said the documents don’t exist (which we know for a fact is complete fucking bullshit), but the CBSA apparently disclosed all of two pages on details on how the Canadian Border Services Agency is using Machine Learning at the border.
Now, if we talk about Facial Recognition in particular, there’s a LOT more information. These documents will require a lot more analysis, but it’s clear that the Canadian Government is more content with just buying off-the-shelf instead of building a solution.
I’m obviously going to have to do a lot more digging, since I don’t think anyone asked CSEC what they’re doing with Machine Learning, and while I’m almost certain that I’ll get a big ol’ “Go fuck yourself” from them when trying to request that info like how I got when I asked how many people from CSEC attended DEFCON, I’m sure that’s where the interesting uses of Machine Learning technologies are being developed for deployment by the RCMP and CSIS are being worked on.
OK, but WHY?!?!?
For people to have an intelligent conversation about something, they have to know what is happening. Right now we could already be stuck in a dystopian future where people can’t go anywhere regardless of whether they leave their phone at home or not because everyone who has a surveillance camera is already running real-time inference on every image that the cameras capture and is indexing the facial recognition results against a massive database. This is very well in the realm of possibility with the cameras that were purchased by the V2010ISU-JIG during the Olympics that just vanished after the games. Furthermore, I would argue that Clearview AI is not even all that special and that anyone can use existing open source tools to do this and that there are already tons of these weird parasitic contractors out to do exactly this. If society has any way to stop it, we need to figure out if and when the conversations happened, who was involved and whether the systems have been deployed. Unfortunately, this is a boring and slow process.
The purpose of this post was to point people to the search interface and hopefully inspire other people who work faster than me to get to this first, since the conversation around these systems needs to happen globally and as usual Canadians will just shrug and say “At least we’re not Americans”.
That’s not good enough, settler.