Stuart Armstrong argues that there are strange benefits of a total surveillance state. According to him, it will drastically reduce crime, reduce cost, prevent (or help recovery from) disasters, boost research, and make our life so much more convenient. This is an interesting and thought provoking read. (And to be clear he is not advocating or defending a total surveillance state; he simply tries to analyse potential benefits.) But I do think he is wrong on several accounts. Especially, I think mostly the ‘happy few’ will benefit from total surveillance.
Stuart Armstrong’s argument
Let’s analyse his argument in a little more detail. Stuart really assumes a total surveillance state where we all live in a panopticon. Everything we do is transparent to everybody else, and is recorded for later reference. He claims this brings the following benefits (which I briefly summarise here, with the main arguments expressed in my own words mostly).
A drastic reduction of crime. In the panopticon, the police is alerted in real time whenever a crime (like a burglary, or an assault) takes place. When responding quickly, they may be able to intervene while the crime is ongoing. With a 100% detection rate, people will no longer commit such crimes. Moreover, all evidence of an ongoing criminal act, or the preparation of such act, will be recorded and can be used later to investigate. This is useful to combat crimes like fraud. It will also prevent child or sex abuse, or abusive working conditions: the evidence exists undeniably, and can be retrieved much later when the abuse is reported after the abused is out of the abusers control.
Cost reduction. With total surveillance in force, and with crime rates plummeting, the police force can be made much smaller. Criminal investigations need far less resources as guilt or innocence would be obvious given the recorded evidence (which contains all little pieces of information that might be relevant).
Also, there would no longer be an arms race as finally reduction of armaments can now be reliably verified. Intelligence agencies can also reliably determine what are genuine threats. “Freed from fear of surprising new weapons, and surprise attacks, countries could safely shrink their militaries.”
Prevent (and quickly recover from) catastrophe and disasters. In a transparent society, pandemics can be detected in their early stages (cf. Google Trends predicting the flu). Affected people are known immediately and can be treated quickly, and specific areas with a high number of infected people can selectively be quarantined. Moreover, disaster recover becomes much easier once the emergency services know exactly who was where in a building that collapsed during an earthquake (this idea is sometimes also referred to as participatory surveillance).
Similarly, surveillance also prevents terrorists from using nukes, or dirt bombs: the bad actors will caught because the whole supply chain is monitored in real time.
Improve research. Total surveillance delivers an enormous, valid, data set describing real human behaviour, in real (not simulated) situations. There is no bias in the selection of the test set, nor is there bias in the measurements. This is a treasure trove for social scientists, economists and epidemiologists to validate their theories. Every action, no matter how inconsequential it may seem, becomes grist for the statistical mill.
Increased convenience. Standard things, like logging in, or even paying at a counter, become obsolete. You no longer need to enter your password: the system already knows it’s you. There is no need to fence of a festival area, to queue at the point-of-sale, or to implement Digital Rights Management. Your credit card will automatically be charged for everything you consume.
On a superficial read, Stuart’s argument appears convincing. But on closer scrutiny, there are issues with almost all of them.
Perfect surveillance does not exist. Stuart’s argument strongly depends on the assumption that the panopticon is perfect: everything is seen in full detail, without error, without omissions. Perfect surveillance is highly unlikely to be ever achievable in practice, though. What’s worse, certain actors may be able tilt the perspective of the panopticon, and influence whether certain raw data is collected or not. Maybe certain people can reliably hide from surveillance (either completely, or every once in a while). Maybe identification of perpetrators is imperfect. People can wear masks to avoid face recognition, or may clone the RFID tag of someone else. If so, these people can commit crimes without ever fearing punishment, or can go shopping for free.
Also errors and mistakes are certainly possible and need to be dealt with. We cannot blindly assume all data is correct. We therefore need ways to verify the data. Is this always possible? And against what cost? Is there any redress if my credit card is charged for a ‘purchase’ I never made?
Data does not equal information. The raw data collected by a surveillance system does not immediately correspond to usable information. To distil this information from the data, the data must be analysed. This analysis is done by algorithms, that base their outcome on a certain interpretation of the data. There are almost always different interpretations for the same observed data point. So the outcome (the information) is by no means unbiased. It depends on the assumptions and interpretations encoded in the algorithm. Who decides on the choice of interpretation? Who can influence this interpretation? Are there bugs or backdoors in these algorithms that allow experts to influence the outcome?
This makes guilt or innocence less obvious than Stuart makes it out to be. As the Talking Heads already sang: “Facts all come with points of view.” Moreover (arguing against cost reduction a bit more) any decrease in spending on the police or the military must be offset against the increase in spending on developing and maintaining such algorithms.
The possibility to influence the interpretation of raw data also allows the generation of disinformation or counter-information to spread FUD (Fear, Uncertainty and Doubt).
In practice, access to information is limited. Not everyone has unrestricted access to all information. People may lack the skills or the resources (they may have old hardware, may not have the necessary software to read the information, or may be connected over a bad network). People with the power to do so, may withhold information, or restrict access to it. Even if you have unfettered access, you still have to know where to look before you can find it. Even worse: you have to know that you have to look in the first place!
Knowledge does not necessarily lead to action. The main problem with Stuart’s analysis is however the following:
- not all information is acted upon,
- not everybody has enough power to act, and
- those that have the power to act may not be willing to.
We all knew that chemical weapons had been used in the Syrian civil war (and from my point of view the question who used it is really irrelevant…). Still it took ages to act, and even then the action was only a halfhearted attempt. According to experts, a clear separation between investment and commercial banks is necessary to prevent a new economic crisis. Still, no fundamental changes to the banking system have been made. Why not? Because banks have tremendous power and resist the change. (See also this earlier post.)
As Aaron Swartz rightly observed:
Transparency […] simply shifts the work from the government to the average citizen, who has neither the time nor the ability to investigate these questions in any detail, let alone do anything about it.
(He has many more interesting things to say about the not-so-usefulness of transparency, for instance the problem that a lot of data collected and stored is wrong…)
Let’s take another simple example. The content industry is perfectly able to ensure that I am charged for all the content I consume. I, however, have very little power to contest a purchase, let alone negotiate the terms of the contract (if, in a surveillance state, there is even such a thing as a purchase or a contract).
It is much harder for an individual to contest a bad decision from government or a business, than the other way around. Even if all the information about the decision is out there in the open. Information is important, but without the power one cannot act.
Coming back to the benefits of surveillance brought forwards by Stuart, I do see his point that it will help to prevent or recover from catastrophes. (Although I am less convinced of the argument that it will prevent terrorism; see the discussion above on the use of chemcial weapons in Syria.) Also, the huge volume of data it delivers will be of tremendous use for socio-economic research (even though it is less extensive and error-free than we might think): it is certainly more extensive and more ‘real’ than the data researchers currently have to make do with. Also, it may make life more convenient in certain cases – if only so that the Chinese waitress already knows that I prefer not to eat, say, pig stomach. Personal preferences are a boon on the net.
However, the main benefits to surveillance and a total transparent society are for those that have the power to act on the information, and use it for their own benefit. In most cases, those will be the government, the large corporations and institutions. Individuals do not have that power. This means that we will only see a reduction in crimes that are considered important enough act upon by those in power. And the inconvenience of having more convenience (sic) squarely lies with (powerless) consumer.
Moreover, as all and everything is known about each and every single individual, we will have lost the ability to hide as a defence mechanism. In other words: transparency disturbs the power imbalance even further.
(P.S.: For Dutch readers, Lynn Berger investigated how transparency became the answer to everything.)