The Sound of Surveillance

The increasing presence of networked devices in our lives has vastly increased the opportunities for harvesting data. But the new frontier in data extraction is based upon the ability of these active devices to listen to us. We are seeing a vast expansion of a new kind of sonic surveillance.

David Beer
5 min readJan 22, 2017

The 1974 film The Conversation leaves us with the image of an increasingly fraught surveillance expert trying to work out how he is being spied upon. Frazzled by paranoia he strips his apartment bare. The fixtures, fittings and even the plaster from the walls are removed in a desperate attempt to extract any secreted bugs or devices. The undiscovered and unsuspected source, as is suggested in a tech-fair featured earlier in the film, is his telephone — the microphone of which can be turned on even when the receiver is in place. The ageing pro is being destabilised by new surveillance tech. This new surveillance is based on existing and embedded technologies rather than added bugs or cameras, making it much harder to detect or remove.

The Conversation suggests a concern with the impending power of audio surveillance. The outcome it points towards is a future of inescapable ears and an associated rising paranoia. When we think about surveillance today we often think about the importance of data, which we imagine as binary code of 0s and 1s, or as text, numbers and visual images — photos, film, data visualisations. We think of the new surveillance powers in visual terms. The panopticon is our reference point for thinking of a world in which we can always be seen. But seeing is not the limit of the data extraction today, it is also in the hearing.

The 40 year old visions provided in The Conversation no longer seem particularly intimidating. The reason for this is simple. We now seem comfortable with surveillance technologies being embedded in our lives, most of the time they go unnoticed or we give them little thought. Surveillance is part of consumption, it’s presented as the pay-off for convenience. The Conversation seems quaint in comparison to recent accounts of the call centre, to pick out an example of a particularly data rich workspace. We now have a vast assemblage of technologies capturing our work, leisure, consumption and even bodily routines. The seeing is escalating, but so too is the hearing. In fact, it seems it is the audio data that is now starting to boom in a way that couldn’t easily have been envisioned.

It would seem that the forces associated with a predictive and automated capitalism are driving sonic data harvesting, which is then used to develop a more rounded and detailed knowledge of our lives. This knowledge can be used to tailor, target, predict and promote. The soundscape is now trackable in ways that previously were not possible. The audio properties of social space can now be interfaced with networked technologies. Soundscapes can be transfered into data, meaning it can be treated like and combined with other data.

Most obviously we have the listening powers of ubiquitous smart phones. Reports that Facebook is listening using the microphones on smart phones caused some reaction (and also some debate about exactly what the microphones are used to do). The suggestion was that the Facebook app listens for targeted advertising opportunities, based on background music or TV and the like. It’s a feature that can be turned off, but it is the possibility of being listened to by the apps that have access to our microphones that is important here. This gives a sense of the scale of potential sonic data extraction and of the pervasive listening infrastructures in which we now live.

In addition we also have devices like the Amazon Echo. These are beginning to obtain some sales momentum, but they point to a future of audio activated devices in the home. Recent media coverage asking what these devices hear and what is extracted gives a sense of the new data being gathered. These devices only start listening when their keyword is used, but they are still listening out for that keyword. It’s not clear exactly what audio is captured. A recent court case involving an Echo may reveal some of what is extracted, if the data is made available as evidence. Clearly the Amazon Echo is opening up new questions about the ethics and scale of sonic surveillance in automated consumer services.

To point to just one other example, the workplace is also rapidly becoming a site of data led sonic surveillance. The new worker badges that listen provide an example of the networking of the sonic spaces of the workplace. These badges Listen out for the tone of voice in customer interactions and the like. These possibilities for sonic analytics expand as part of broader trends in workplace tracking.

The Conversation may now seem dated, limiting its ability to provoke our fear of surveillance. The home-space and workplace have now been networked for some time, we are used to it. Plus it’s vision pales into almost insignificance when we reflect on how the whole of the social world is now embedded with microphones. As convergence led to a far greater presence of cameras, so too the embedding of phones has led to digital and networked microphones being everywhere. Despite these changes the 1974 film does still suggest something. It suggests that when we are confronted with bemusing listening environments we, like the surveillance expert, may end up frazzled with paranoia unable to find and strip these embedded bugs from our everyday lives. The desperate efforts of Gene Hackman scrabbling to find out the source of his leak may be an image that somehow captures a future in which sonic data becomes the new frontier in data based surveillance and tracking.



David Beer

Professor of Sociology at the University of York. His most recent book is The Tensions of Algorithmic Thinking.