Submitted by Bryce Laliberte

We are living through the last gasps of an era of anonymity and freedom, one in which the encumbrances of surveillance are surrounding us on all sides. Although it is already well-known the ways in which private companies gather myriads of data on everyone they can, from credit card scores to legal records to social media, these methods are at least difficult to access and closed to casual perusal. However, we are on the eve of a revolution in surveillance, powered by machine learning, drones, apps, and public accessibility.

As it becomes easy, even trivial, to generate – and access – public surveillance records to find what we are looking for, we will all be forced to adapt to a regime of popular will in which any and every of our public actions can and will be scrutinized by anyone with an internet connection. This is the Era of Surveillance Twilight, before public records encompassing all actions occurring in the public are uploaded to the internet for easy search and observation by anyone with the least interest in doing so.

First, what must be understood about the production of surveillance records is that they are useless without a means to detect and identify meaningful events. A 24/7 record of people interacting in a public space will, if there is a crime – or some other action – committed, successfully record that crime occurring. Unless someone already knows about the event in question, searching through that footage will require a human to manually and laboriously look through all the footage. This is a method sufficient for making recording worth our time, but this means that many noteworthy and interesting events are overlooked. Moreover, the systematic search through footage of multiple recordings is prohibitively time-consuming. Trends and patterns observed of behavior, unless an observer already has some other reason to look for it, go unnoticed and, in essence, unobserved.

Further, the overwhelming bulk of recorded footage is inaccessible by the public. One reason for this is that it is not, in the present, in the interest of most parties making recordings to do so. A 24/7 recording of a bank might just be used by a would-be bank robber to identify patterns in worker behavior they might use to their advantage. The bulk of recordings provided for public consumption are either for hobby or entertainment purposes, with the prohibitive time-cost associated with searching through public footage reducing demand to a negligible amount.

There are, however, noteworthy exceptions to this general feature of surveillance records. For example, the Ring Video Doorbell system recently being offered to consumers offers a relatively simple motion detection system, providing alerts and recordings to customers of the motion events it detected. Obviously, the average person would not care to obsessively watch surveillance footage of their home 24/7, but the detection and capture of motion events at least reduces the footage needing to be looked at to those which are short and most likely noteworthy. In combination with an app offered by Ring that allows included neighbors to look through footage shared from the doorbell camera and police to access recordings, we have the beginning of what might be called a Publicly Accessible Surveillance System. In other words, it is a primitive – and very real – panopticon.

What makes a Publicly Accessible Surveillance System (PASS) significant is that it provides the tools to realistically document every behavior occurring in public and, most importantly, make those public behaviors easily searchable by the public. Motion detection is a simple but limited means, as the categorization of motion events still requires manually observing the footage by a human, who will then determine whether the event in question is innocuous, criminal, or interesting for some other reason. In other words, the detection systems currently available to the public are primitive, but future developments in surveillance technology – especially through machine learning – will radically expand the scope of surveillance systems and, by extension, their use and consumption by the public.

At present, facial recognition technologies have reached a point of effectiveness that they are being adopted by China as a means of identification. Rather than entering a password or swiping a card to access an account, your account used for various daily necessities, from using the subway to accessing your bank account to making purchases, is being attached directly to your face. The systems are effective to the point of monitoring the location of everyone in public and their whereabouts. In China at least, where you come from and where you go is practically impossible to keep a secret from the government, at least in highly populated public spaces. Future developments to their comprehensive surveillance system will likely include drones equipped with cameras to monitor areas not already under surveillance by immobile cameras. China’s development of public surveillance systems is meaningful for those elsewhere, as their efforts to invest in and develop the technology will result in the technologies associated with comprehensive surveillance falling in price. As the price of surveillance technology falls, it will become increasingly in the reach of private consumers who will be interested in protecting themselves in their own homes and neighborhoods.

In the United States, facial recognition technology is not yet available to the public in a form easily accessible for consumers, but that development is only a few advancements and price reductions away. For example, it is an easy and obvious improvement to currently available consumer surveillance systems to introduce facial recognition, at least so that individuals can be categorized and their whereabouts might be consistently updated when they appear in footage. The next step, that of identifying individuals based on facial recognition, only lacks a centralized database accessible by private companies and individuals, but then that only represents a business opportunity. If strangers are walking around a neighborhood, consumers will want to know who those people are, if they live nearby, if they have a criminal record, if they are observed engaging in some legitimate reason to be there, e.g. visiting a neighbor or store. Although we shirk at the idea of losing our personal anonymity in public, we are generally willing to sacrifice others’ anonymity for the sake of protecting ourselves or, more meaningfully, our children.

Likewise, when it comes to children, parents will be interested in knowing when their child comes and goes and where they are. Future updates to the Publicly Accessible Surveillance System enabled by systems such as the Ring Video Doorbell might include using facial recognition technology to observe as that child goes around the neighborhood, being recorded by neighbors’ video doorbells. Likewise, as drone technology improves and falls in price, it is conceivable upper and upper-middle class neighborhoods – especially those with vigilant neighborhood associations – will subscribe to a system involving drones that fly around the neighborhood, providing surveillance on a mobile and 24/7 basis.

An important development in surveillance technology will be making recorded footage searchable. That is, it will eventually be possible to enter someone’s name in a search bar, and a machine learning system will automatically search through footage of various searches to document when and where that person has been observed by a Publicly Accessible Surveillance System. As searchability of surveillance footage improves, it will synergize with the falling cost of surveillance technologies, creating a virtuous cycle of expanding surveillance alongside increasing granularity in who – and what – people might analyze surveillance for. Not only will the square footage of surveillance increase by magnitudes, but the utility of that surveillance will also increase at the same time.

The possibilities of searching surveillance records do not end with identifying individuals and their whereabouts, NSA-style. Retailers will drive the push for capacities to identify specific objects, technology that will doubtlessly end up being used outside retail contexts. Surveillance AI will also be capable of determining whether individuals have weapons on them based on their gait. With time, the granularity will improve to the point AI will be capable of identifying specific behaviors, from the innocuous such as a father and son playing catch to the criminal such as an individual casing a house. All these specific behaviors will be categorized with keywords making them all searchable, so that one might not only make a search for the location of a specific individual, but for all instances of certain behaviors such as individuals exchanging money and drugs.

It is likely that comprehensive surveillance associated with local police departments will provide the most detailed recordings of public behavior, but Publicly Accessible Surveillance Systems will be only shortly behind in scope and ease of use. The result will be a society in which every public behavior can be easily searched by any individual curious enough to look for it, with anyone having the ability to use search engines akin to Google to find out everywhere a person has been short of them taking extensive measures to counter surveillance. Crime will likely be deterred at the same time new crimes become possible. All of this as the end result of people simply desiring to secure their home environment using surveillance as a means in a society quickly shifting to a low trust designation with little social cohesion. Everyone will become watchable by everyone, in many ways turning the surveillance state inside out. If you are a public figure your actions will receive tenfold scrutiny by citizens, and only if you’re a nobody perhaps you will avoid the intense gaze of the collective.

That is why this period of time shall be known as the Era of Surveillance Twilight. Although surveillance exists, it is in the simplest and most primitive form, barely capable of serving the purposes we now task it. Developments in machine learning and the private centralization of data for identification will increase the utility of surveillance tenfold, leading to a society of surveillance radically impacting society at the most basic level of public behavior. Surveillance technology is only beginning to inch to its final form, realizing the ultimate panopticon permitting everyone to, at last, Watch the Watchers for the simple reason that everyone will be equally Watched.