CCTV spooks to be spied upon by gaze-tracking webcams

A reader writes, "You couldn't make this up: Cameras are being turned on the people paid to watch CCTV streams, to note which bits of surveillance footage they didn't see. The system developed in Turkey uses webcams to track a person's eye movements and can then produce an edited reel of footage that they didn't see at the end of their shift."
Privacy campaigners may enjoy the irony if the gaze-tracking system comes to be regarded as intrusive by CCTV operators - who could fear that employers will use it to dispense with their services if they consistently miss too much on-screen skulduggery.
Eyeball spy turns the tables on Big Brother


  1. Robert Anton Wilson pointed out the logical conclusion to the problem of surveillance of citizen in TSOG. Each group set to watch will have it own watchers, who in turn will be watched, and watched and watched, till everyone is spying on each other and people starve to death through lack of short order cooks.

  2. This seems impossible, unless you have one person watching every single monitor, and even then it doesn’t work.

    If you have one person watching four monitors (in real time, of course), then a total of four hours of footage is shown to them over the course of one hour. That means, no matter where they look, three hours of footage will not be watched. Making them watch the three hours at the end of their shift doesn’t work, because then they’ll be three hours behind the current stream.

    If you have one person watching one monitor, if their gaze wanders for 10 minutes in an hour, they have ten minutes of watching to catch up on at the end, so they will keep falling further behind the current stream.

    …Ok, I guess they probably have an algorithm that decides “if you look at this monitor for at least one second every five seconds, we’ll count that as having watched the monitor for the entire five seconds,” but it’s still a little silly, logically.

    What’s interesting to me is that it means that there really are people watching through those CCTV cameras at all times. I always just assumed that most of them were on autopilot — recording, but no one would watch them unless there were a need to look over the tapes. Scary.

  3. Samsam, I still don’t think there are people watching live feeds from most cctv cameras. It’s simply a question of logistics. The amount of employees you’d need to cover even a small fraction of cctv feeds is just silly. Even if you tried, you wouldn’t cover any significant amount of feeds, and thus the perceived advantage of having someone actually watching live feeds will be negligible, meaning that a competitor who hires fewer live-stream watchers will outperform you.

    Yay capitalism!

  4. There was a case here a couple of years ago where an entire shift of CCTV ops went bad and started pointing them into women’s bedrooms, looking down dresses and so on. They were caught, sacked, and jailed, including the shift supervisor, on the strength of the surveillance images of the CCTV control-room. Let’s see if I can find a link… hmm, I think this is it, although it doesn’t mention the “cctv of cctv operators” angle. Oh wait, here we are:

    Mr Davies said the camera monitoring the CCTV control room showed Judge turning down the lights as the film was shown on a large screen.

    He added: “Mr Welsh is seen to be in an avid and animated mood. He is seen to point to the screen and raise his hands to his eyes in a gesture to suggest binoculars.”

    (emphasis mine.)

  5. Yup we now know everyone is being watched, even the people who you don’t think are being watched. Oh dear, I’m probably being watched right now!

  6. I thought it had been proven without a doubt by now: the only people ever watched by CCTV-watchers with more than one monitor in front of them are either sexually interesting to the watchers or young black men (who seem to be universally suspicious to idiots with no understanding of math, history, OR psychology)

  7. nobody has directly pointed out the impending overlord implications here; since it is software that is “automagically” making this “greatest misses” reel for the operators at the end of their shift. that leads me to speculate that this software will improve over time to do the bulk of the watching for the humans. then the job of the operator is to only review what the machine thinks is significant. not long after some cheapskate employer will do away with the operators all together. then our lives will be 200% observed (at least two cameras for every location) by automated systems that determine if we are naughty or nice. not long after debt collectors begin tapping in to the facial recognition systems to locate deadbeats. advertisers take targeted marketing to a new extreme. anyone questionable gets arrested. the machines keep us safe, from everything, most of all from ourselves. we all stare in amazement at the pre-recorded images of suicide bomber’s faces moments before they blow themselves up; and are puzzled as to why the machines can not prevent serious crime while being so good at punishing those that mostly abide by the law.

  8. There’s a simple fix to all of this: two-way visuals. Above the operator’s video screens, put a camera. Above the cameras out in the streets, place a video screen. Now you’ll always know if somebody is watching you and what they are doing.

    While you’re at it, do what a lot of retailers with “scare cameras” do and, on a screen by the camera, show what’s being recorded. Then it’s clear if somebody is pointing the camera where they shouldn’t be.

  9. #4 I was just about to say that this sounds like the beginning of a horrible real-life version of A Scanner Darkly.

  10. i’d be more interested in seeing what they were watching. nuts to the guy who missed the drug deal- what about the jackass paying too much attention to our friend the living statue?

  11. the missed=footage software should also have a value-assigning system and the CCTV attendant should have a shock collar that punishes him for missing important things.

  12. آدمك‌ ، آدم‌ خرد ، كوتوله‌ ، گور زاد

  13. Takuan @21.

    آدمك‌ ، آدم‌ خرد ، كوتوله‌ ، گور زاد

    My Farsi is a little rusty, and Google Translate won’t even give me a hint. Something to do with Homunculus by Hideo Yamamoto?

    Would you care to share with us?

  14. Takuan @24:

    I see what you did there! Because I did the same thing… Also I missed that you were actually replying to OrangeBag @20

    At least now I know what homunculus means. I will use it at work today to further obfuscate my oratorical dissertations.

Comments are closed.