Surveillance

The Panopticon is a theoretical prison imagined by Jeremy Bentham in the 18th Century. The Panopticon features a tower inside a ring-shaped prison. The tower has windows facing outwards at the prison, and all of the cells have windows facing the tower. This means, in theory, a guard could (from the tower) view any of the rooms and see what any prisoner is doing. Of course, one guard could not watch every single room at the same time, but the Panopticon is designed in such a way that the prisoners can not see the guard. Therefore, at all times, the guard COULD be watching a prisoner without them knowing.

With technological developments, surveillance has changed. Now a guard does not need to stand watch; cameras, smartphones, motion sensors and even computer software can track members of society. When we are asked to consider surveillance, we think of CCTV cameras. However, even our own devices are now capable of watching us. Our phones respond to facial recognition and store information about our appearance, our SmartHome devices such as Amazon Echo and Google Home listen to our conversations and monitor our tastes. We live in a data capture society where every keystroke we input onto a computer is recorded.

This 2013 list from The Argus documented the location of all CCTV cameras in Brighton at the time.

Whilst there is a website for speed cameras in Brighton, I could not find an up to date map of Brighton CCTV cameras. I went to the Sainsburys at the end of my road and attempted to map the layout of the cameras in the store, until an Assistant Manager asked me to delete the photos and leave. I find it interesting / worrying that they’re allowed to watch us, but we can’t watch them.

A fascinating video uploaded to YouTube by Alexey Kadinsky demonstrates how machines (robots) can ‘read’. I would be cautious about using the terms robots and reading. Reading involves some form of understanding and comprehension, and these cameras seem to just track moving objects. Robot also implies some form of creative, autonomous intelligence, and seems a little bit too Sci-Fi for my liking. The video, nonetheless, is useful for visualising the way in which technology can be used in surveillance further than just recording video footage to be reviewed by a human being.

Hiding from machines

As with the majority of technology uses, the history lies in warfare. During World War 1, U-Boats caused a huge problem for British ships. A British artist, Edward Wadsworth, created ‘dazzle’ designs on ships in order to limit the effectiveness of U-Boat tracking systems. These ‘Dazzle Ships’ confused U-Boats when calculating distance and speed, making it harder for the ships to be hit. This method was less about ‘hiding’ and more about confusion. Perhaps confusing technology is just as useful as hiding from it?

Dazzle Ships in Dry Dock – Edward Wadsworth (1919)

A piece published by DIS Magazine, written by Adam Harvey, offered these four steps to avoid being detected by facial recognition software:

1. Avoid enhancers
       They amplify key facial features.

2. Partially obscure the nosebridge area
       The region where the nose, eyes, and forehead intersect is a key facial feature.

3. Partially obscure the ocular region
       The position and darkness of eyes is a key facial feature.

4. Remain inconspicuous
       For camouflage to function, it must not be perceived as a mask or disguise.

Harvey’s work is in response to (amongst other things) the introduction of Janus in America: a program which collects images of faces from social media and videos, which he argues is a “threat to privacy”. Harvey proposes using dazzle techniques in configuring your appearance to confuse this software and limit its ability to recognise your face. Methods include removing symmetry in your face, inverting normal facial tones and hiding the nose.

DeepDream

The software featured in the video above involves machines identifying patterns in images, and classifying them as a result of these patterns. The ImageNet Project is a large database of ‘visual recognition’ images, working on developing accuracy of classifications and identifications within images. This involves training machines to notice patterns in images and assign them a label. Once this label is assigned, the software can label these patterns in other images.

DeepDream is a piece of software designed by Alexander Mordvintsev, a Google engineer. This software identifies patterns within images and, instead of just classifying them, it enhances these patterns. The hyper-processed images create a hallucinogenic, or dreamlike effect – hence the name DeepDream.

Using technology to classify and identify patterns in images is one thing, but technology can also be used to create realistic images. Terence Broad used an auto-encoder (a type of neural network) to recreate the film Blade Runner. The auto-encoder is fed a data sample (from the movie) and then reconstructs this data sample. Broad raises an important point at this stage, which is to avoid anthropomorphising this software. The anthropomorphism of software and technology is a fascinating intellectual, moral and philosophical debate. In 2002, Murano suggests that “anthropomorphism at the user interface in the context of online systems usage is more effective than a non- anthropomorphic method of feedback” (See resources).

This Person Does Not Exist is a website which features images of people who simply do not exist. The images on the site are created by Generative Adversarial Networks. Whilst the site may appear to be harmless exploration of the capabilities of deep-learning software, the creator, Phillip Wang, wants to raise awareness of potential sinister uses for this sort of software. This highly realistic approach could be used to impose faces on existing images or videos in order to spread fake news, propaganda or for other purposes.

As with my other blog posts, I have noticed that deep-learning, image classification and image construction programs could be used either for security/art/science or for deception/crime/invasion of privacy. The reality is that, whilst new technologies provide new ways of aiding society, they also provide new ways of endangering society. Much like the development of any tool, the value and purpose of that tool lies in the hands of the person wielding it.

Resources

Comments are closed.

Blog at WordPress.com.

Up ↑

Design a site like this with WordPress.com
Get started