Last Updated on May 4, 2024 by Karl Thompson
The advent of digital society has powerful implications for social control, surveillance, crime and social policy.
Digital societies are those in which digital technologies are today integrated into daily life. They have powerfully changed the way human societies are governed.
Digital technologies provide massive volumes of data which are analysed and used to inform policing practices and social control policies more generally.
The police use big Data to construct gang databases for selective surveillance, and in sentencing decision making.
This post explores how crime control is changing with the rise of Smart Cities, focusing on predictive policing as part of this. It also applies several social theories to understanding social control in digital society and examines criticisms.
Social Control in Smart Cities
Smart Cities link digital and physical infrastructure to enable social ordering.
They are often presented as being desirable places to lift, with some accounts being utopian. However, they also allow for huge volumes of data to be collected and used to control citizens.
According to Laufs et al (2020) there are three layers of technologies within Smart Cities…
- The sensor layer which are data collection units. Most obviously cameras, but this also includes facial recognition software.
- The network layer – the infrastructure to analyse and aggregate data.
- The actuator layer – alerts which inform actors to act on the basis of data collected. Fully automated machines or human beings will do this analysis.
Thus in Smart Cities hardware, software and human beings are all integrated into a crime control network.
How Smart Cities change Crime
Smart cities open up new opportunities for certain actors to commit crime, and make new types of crime possible:
- There is more potential for domestic state crimes to take place. The State now has more data on more people than ever in human history.
- There is more scope for international state crime to take place. Transnational Corporations have greater access to public data.
- Corporations such as Meta harvest much of our data. Thus there is more potential for Corporate Crime.
- There is more potential for organised crime groups to hack state or corporate data.
- There is more potential for cyber-terrorism.
Digital Cities have multiple attack vectors for criminals:
- Weak software and password security is one way in.
- Poor maintenance of out of date software systems.
- Cascade effects with higher levels of integration.
- Criminals may exploit human error and disgruntled employees as a way in.
Examples of damaging cyber-crimes:
- The 2019 Suxnet Worm attack on an Iranian nuclear facility
- The ‘WannaCry’ ransomware attack which damaged the NHS’ patient databases in 2017.
- Russia’s cyber attack on Estonia in 2007 which targeted banking, news providers and voting systems.
Potential crimes we may see in Smart Cities include:
- Taking control of traffic networks and vehicles.
- Attacks on Smart Buildings – control of lifts, lighting, heating.
- Healthcare emergency and response systems
- Falsifying payments through smart metres, for example.
Implications for security and surveillance
We are already seeing the increased use of Body Worn Cameras by the police. We also see more Unmanned Aerial Vehicles (Drones), and the increasing use of GPS tracking to control crime.
Technology in Smart Cities can prevent crime in the following ways:
- Detecting crimes through cameras and other surveillance tools.
- Authenticating at secure entrances to only allow authorised persons to enter buildings.
- Identifying actual criminals through facial recognition software
- Profiling the types of people who may have committed a crime, using existing data.
- Tracking objects through GPS, which can help prevent theft or find stolen goods.
We now have an increasing array of sensors beyond cameras. These include acoustic sensors (to detect gunshots for example), and atmospheric sensors to detect potential harmful substances. Cameras have also evolved to track gait analysis and facial features. There is even ‘sentiment’ analysis.
In terms of analysis the police are using more and more data in certain cases. Digital Forensics is a growing field, used especially in child abuse cases. This might sometimes involve hours and hours of searching through all phone records or browsing history.
Predictive Policing
Predictive policing is where police forces use existing data to predict patterns in future offending.
Four ways in which policing has become predictive:
- Spatio-temporal – where and when are crimes more likely to occur..?
- Predicting offenders – what types of people are more likely to commit crimes in the future..?
- Identifying what types of offence are more likely to occur when and where and who is more likely to be committing the offence.
- Identifying victims.
Predictive policing has increasing amounts in common with actuarialism (insurance). This is where the police calculate the risks of offending taking place based on past data.
This means that algorithms based on past data are increasingly determining where police resources should be deployed.
One company which worked with several police departments in America was PredPol. You can view a slide show of its presentation here. It outlines how its algorithms will show police at the beginning of a shift where the crime hotspots are in their areas.
NB evidence suggests this software is very inacurate!
Criticisms of predictive policing
Predictive Policing algorithms are not neutral. Old biases will feed into them. For example, historically the criminal justice system has been biased against young black men. This means any algorithm used to predict future offending may exaggerate the chances that any young black man could be a potential offender. In 2021 The Los Angeles Police Department scrapped their predictive policing programme because of this.
These biases hold true for a range of characteristics: those from the lower social classes, in poverty, from inner-city estates, for example.
Where algorithms are concerned there may be exponential combination effects when calculating risks of offending. Predictive Policing software is much more likely to flag someone with multiple ‘criminal risk features’.
Historically there is more data collected on the powerless than the wealthy elite. Thus, predictive policing models are more likely to catch the powerless, not the wealthy elite who have always committed crime but never got caught doing it. (Which is the case if you are coming from a Marxist perspective on crime!).
Theories of surveillance applied to Digital Society
Foucault’s classic theory of the Panopticon model of surveillance doesn’t work here. According to Foucault the general population are more likely to regulate their own behaviour because they know that a central authority is observing them.
This panopticon model doesn’t capture the way surveillance in the Digital Society works. Surveillance today works by categorising people into those who need to be surveilled more and those who don’t. This is the case with predictive policing, for example. Here there are huge numbers of people: mainly middle class, older and white who simply aren’t in the police models as potential criminals. However, if you’re young, black, male and poor, the surveillance net is more likely to catch you.
Similarly, airport surveillance makes life easier for most of us, by making security checks quicker. However if you have a male muslim profile from a certain country, you are more likely to be checked.
Mathieson’s model of synoptic surveillance works better in some respects. According to Mathieson we are all observing each other and this is where control comes from. This includes the relatively powerless observing the powerful. The case study of George Floyd is an example of this: where ordinary citizens filmed the police murdering him.
In order to fully understand surveillance today we need to understand that we exist within a ‘surveillant assemblage’. This is where the state, the police, and private companies are all involved together in a social control network.
This is easy to see if you consider how many surveillance technologies are created by private companies. However they are then paid for by the state and used day to day by the police. Facial recognition software is a great example of this.
Actor Network Theory may be useful in helping to understand social control through surveillance in digital society. Actor Network Theory holds that technology acts like an agent. This means that while people shape technology, technology also shapes people and their environments.
Relevance to A-level Sociology
This material is most obviously relevant to the module in crime and deviance. It is also relevant to social theory, research methods and social policy!
Relevance to social theory
This is extremely relevant to postmodernism and approaches to social control in a post, or late-modern society….
Following Swift (2005) we live in an age of informational capitalism. In this age the distinction between online and offline worlds is blurred. Indeed, to be a citizen in the fullest sense of the word we need to be online.
So it is unlikely we are going to see masses of people go offline anytime soon. Thus more and more people are going to be subject to social control in smart cities. More and more people are going to be subject to predictive policing. These shifts in social control are relevant to everyone!
Relevance to research methods…
The use of Big Data in policing policy illustrates a shift in the way research methods relates to social policy…
Big Data has brought two epistemological shifts that have fundamentally changed the way we understand the world.
- We are now less reliant on representative sampling. Big Data means we are more able to get relevant data for an entire population under study.
- Issues of causality become less relevant. Correlations between different data points become more relevant.
The implications for social policy are that speculative knowledge gains more status. Policy makers make data on the basis of correlations between the data points we have available based on digital technologies.
Digital data analysis may appear neutral, but it is always biassed.
Sources
Liebling et al (2023) The Oxford Handbook of Criminology
