Is it technically and logistically feasible to watch everyone all of the time?

In My Humble Opinion
1

Cameras are now extremely cheap. Even surveillance cameras with motors for pan/tilt/focus/zoom and decent lenses are under a grand. The network to interconnect them has also plummeted in price, and it is now possible to have networks with thousands or millions of digital surveillance cameras sharing the same set of wiring. (you’d use a few layers of routers and switches and data compression and optical fibers for the main trunk lines)

None of this was technically possible, say, 25 years ago.

GPS is now ubiquitous and cheap, and, again, it would be practical for every single citizen to be forced to carry a phone or something with firmware that reports the position of that phone every minute or so via wireless data.

I’m not saying that the total surveillance state is a good idea. Quite the contrary. But, could it be done?

Let’s say you had 1 in every 10 Americans hired to watch the other 9. This is the kind of job almost anyone could do. Each person would have 9 people they are “assigned” to, and they would have to write up a report that would be full of automated keywords and macros and stuff to describe what each of the people they are watching is doing.

Most people are extremely repetitive in their daily habits, and spend a lot of time in the same places. So, I think this would be a pretty easy task. The surveillance worker would be able to fastforward past activities that aren’t likely to be seditious - every hour the worker spends in his cubicle working, or browsing time wasting websites on a government approved whitelist, or eating, etc is time you can skip through. Software could automatically flag regions of the surveillance records that actually have conversations and such that might contain suspicious activity.

Then, of course, you have to watch the watchers. So for every watcher, you also need someone to watch them. Perhaps 1 watcher watcher for every 10 watchers. And so on in an infinite series, until you get down to a ring at the top of the pyamid of like 1 person who watches someone else who watches a third person or something.

Could it be done? I mean, is there enough material resources to actually build all these cameras and hire all these people and train them and so on? How would you respond in an efficient manner when someone drops off the grid? Dogs and troops and helicopters are expensive. Can’t deploy them every time.

2

At a ration of 1 watcher for every 10 watched (including watchers of watchers etc), you would need over 35 million watchers to watch the entire population of 320 million Americans.

3

Labor force participation rate is 62%. Being a watcher is a job virtually anyone could do - you just need to be able to describe what a person is doing in a series of text reports. You would need to be able to see, understand english, and have basic literacy - I suspect you could write translation software so that people who are barely able to write and who are used to “text-speak” from phones can write usable reports.

There would have to be overlap. Some kind of rotating schedule where shifts of watchers check other watched individuals to make sure the first set didn’t miss anything. You could test your watcher pool’s writing skills, literacy, observation skills, etc and have each one assigned a job grade. (and you have your A grade watchers watching the people who are most suspicious)

4

You could store a day’s tapes, and then have a trained watcher review them at fast-forward speed. You could also use computer image-recognition shortcuts to abbreviate a typical day’s surveillance tape down to the most relevant hour or two.

5

Seems like I remember that the pit falls of this kind of thing are very bad and make peoples kinda unhappy.

Thought experiment? How depressing.

Raw material possible question?

If 35 million is correct, how you gonna get 35 million to agree to do it and be conscientious about it? 35M ‘people?’ People and all their … er… ah … !!!

PS:??? Restrooms, bedrooms pools while the kids are at school and many more problems? If there can be one full minute of real privacy, then there will be 34M making use of that…, Probably in a not good way.

6

I agree with GusNSpot.

7

It could be done with computers doing most of the watching, sorting most autonomously just referring a few for human viewing for whatever was the standard for that.

8

It wouldn’t necessarily need massive centralised control or computing power either. The surveillance devices themselves could be parts of a distributed computing/storage solution.

9

Let’s start with Shagnasty’s back of the envelop analysis in another thread. He’s assuming coverage of only 10% of the US, less labor (no reports and not having to track by individual), and doesn’t really cover extra cameras to cover every room and space inside buildings. He still comes up with 8 Trillion to implement with annual operating expenses on the order of 6 Trillion. That’s for 10% of the country. Just using the same rate we’re talking 60 trillion dollars without the extras above. US GNI at PPP in 2013 was 16.99 Trillion. Gross World Product (PPP) for 2013 was US$87.25 trillion. We’re talking around 92% of the total economic output of the world just to implement the system in the US without the extra costs.

The let’s looks at other issues that don’t figure in to Shagnasty’s network. Huge chunks of the US don’t have effective cell coverage (in some cases analog let alone data). Now we need to massively extend the cell network, and run power to this omnipresent camera network. Then we need phones for the people who don’t currently have one. GPS signals get blocked relatively easily and phone batteries die so there’s still significant loopholes to lose your tracker.

Do we even need to get to the possibility of an insurgency attacking that system via network or physical attacks?

10

You could serve most of the purpose of “what everyone all of the time” by creating a “watch anyone, anytime” system. That is, it’s not guaranteed that a given person is being watched at a given time, but there is no guarantee that they aren’t being watched (and a significant possibility that they are being watched).

I think the data analysis problem is the easiest to solve, and could drive the “significant possibility” up towards 100%. Rather than assigning watchers, make the feeds publicly available, and create an open market in event classification. If you’re watching for particular behaviors or events, offering rewards for identifying those events will drive both technology solutions and brute-force crowd-sourcing.

If you want everything described, image identification + meta-data will do a lot of it (time, place, person, posture). Anything unclassified in this way (person sitting at computer by themselves, person asleep, and person in conversation with other person will all be automatically tagged) can be flagged with an associated micro-payment for watching and classifying.

Simple watchdog algorithms can look for suspicious behavior such as Person Y doing lots of classification of Person X. With a large crowd of classifiers you don’t need to worry about someone hiding their activity by suborning the watchers. So long as only one country is doing this, you could have many times as many part-time watchers as you have people being watched.

Getting 100% coverage of cameras is a little trickier. Rather than trying to cover every space, you’d need fewer cameras to cover every person. A mandatory wearable camera, plus a couple of indoor drones and a couple of outdoor drones would mean that you’d only need five or six active cameras for each person (and a big industry in maintaining the cameras).

Otherwise you’d need to scale - lots of high-quality coverage for crowded indoor spaces, lower resolution single-angle coverage for open countryside. Everyone would always be watched, but they could limit how closely they were being watched by going to particular places.

11

Your comments are really clever. Everything except the drones following everyone sound like major improvements to my napkin sketch. A distributed network where no one can be certain they aren’t being watched is a lot better idea than a setup where there’s 1 or 2 specific “watchers” who are watching you. I suspect the chief vulnerability of the system would be that the watchers themselves are corrupt. For instance, we know that when the FBI of the 50s, 60s, and 70s had this power, it blatantly abused it to surveil people at the whim of J Edgar.

You definitely want the watchers to get watched so this doesn’t happen, and to make sure they are actually doing their jobs and not just wasting resources, and so on.

The reason the drones are a bad idea is that quadcopter drones require enormous amounts of power to fly and are very noisy and have limited flight time. Just not practical - small cheap microcameras everywhere would be better.

It has occurred to me that in a society like this, you might want to create places where people would reasonably think they are not being watched, but in actuality this is where you install your best cameras and have your most skilled watchers covering.

12

Just what are you watching for?

13

You’re right, I was being wildly unrealistic about the drones. I guess I was extrapolating into a near future where drones were omnipresent, but there’s no Moore’s law for battery power. I wonder if you could manage a scheme where whenever you left micro-camera coverage zones, a long range drone started “following” you - where “follow” just means devoting attention. So you could have, say, thirty or forty long range drones covering a thinly populated area, rather than micro-cameras in trees.

I like the idea too of using apparent lack of coverage as a tool. If people incorrectly think that they can escape the system easily, they’re less likely to be imaginative about escaping the system. Someone who believes there is no camera in the park is less likely to take a boat into the middle of the bay to have a conversation.

I think though that if drones are impractical (and I concede that they are) then the logistic limit of the system is with cameras rather than watchers. Technology and innovative crowdsourcing can at least in principle watch all of the footage. Gathering the footage requires a camera with line of sight to every point in a country, with a margin for temporary obstacles. That’s a lot of cameras.

14

Yeah, if I knew I was being watched, I might choose to do some pretty nasty stuff, all legal but nasty.

15

All crime. You could in principle create a society where no one commits a crime and gets away with it, people are punished immediately for their crimes, and effective measures to reduce recurrence are in use. Since there are surveillance records of everything, the innocent would have a lot less to fear.

16

…

17

Oh. Well, ok, you can fly drone aircraft a heck of a lot more cheaply in open air than in indoor spaces. Google for “google wing” for an example of a drone design that is power efficient enough to stay up for several hours. Yeah, for rural areas, you’d have to quietly deploy drones with infrared and visible light cameras to follow people who are walking around in the woods, etc.

A simpler method than trying to get microcameras and microphones into the trees is that whenever any 2 people are meeting in the woods, you could just offer both of them a small reward if they betray the other and carry a device with a microphone with them.

18

I think these are all excellent points, and basically show that such a program of surveillance would not only be a totally and completely unaffordable government program, and would also completely destroy the American economy.

The logic of spending (let’s be generous here) $4 trillion or more (which is more than double all current government expenditures and 25% of the entire GDP) for this surveillance is self-destructive to the point of being impossible.

19

You need to add vast amounts of security software and monitoring if you want to make sure that the data is kept from prying eyes. And then people to monitor that software and investigate attacks and break-ins. What are your data retention and auditing requirements?

20

Actually, letting prying eyes see everything is kind of the point…