I would like to acknowledge the financial support of the Independent Social Research Foundation, who are funding the research project (Re-Thinking Freedom in a Neuroliberal Age) that this blog is associated with: http://www.isrf.org/about/fellows-and-projects/mark-whitehead/
Geography and Nudges: Real, Virtual and Metaphorical
As a geographer with an active interest in the world of Behavioural Public Policy and nudges I often have to explain the links between my discipline and these areas of research. Recently, however, I have noticed a shift in discussions about certain forms of Behavioural Public Policy towards geographical concerns. In some instances, this shift is, admittedly metaphorical. A case in point emerges from Cass Sunstein’s recent volume On Freedom (Sunstein 2019). In this book Sunstein utilises the notion of “navigability” to explain the value of nudges. According to Sunstein, it is precisely because it is so difficult to navigate an optimal decision-making path through our world (with temptations, sludge (the use of nudges for more pernicious ends), confusion, obfuscation, and exploitatively designed choice environments everywhere) that we need nudges to act like our behavioural Sat Nav. In other words, nudges can be summed us as routing suggestions: it is not that you have to go this way, but it is going to best for you if you do (although for an alternative discussion of the behavioural economics of Sat Navs you may want to read Rory Sutherlands recent book Alchemy (2019) – Sat Navs’ flaws appear to emanate from their rigid commitment to logical calculations of distance and time, which cannot account for the psycho-logics of human mobility needs).
Beyond the metaphorical realm, however, geography appears to be becoming an increasingly important consideration within emerging forms of hyper-nudges. According to Yeung hyper-nudges are nudges that fuse together behavioural and data sciences (Yeung 2016). Yeung observes that:
‘Hypernudging relies on highlighting algorithmically determined correlations between data items within data sets that would not otherwise be observable through human cognition alone (or even with standard computing support (Shaw 2014)) thereby conferring ‘salience’ on the highlighted data patterns, operating through the technique of ‘priming’, dynamically configuring the user’s informational choice context in ways intentionally designed to influence her decisions’ (2016: page 8)
Interestingly the notion of hyper-nudging takes its name from the scaling-up of nudges through social media networks and smart technologies to deliver population level effects. But, as this quote from Yeung reveals, hyper-nudging can also be thought of as micro-nudging—to the extent that it involves the delivery of personally salient nudges to individual in real time (seeDow Schull (2016)). To these ends hyper-nudges would include Facebook’s Voter Megaphone Project (through which voting in elections in promoted on the basis of indicating who in social networks has already cast their vote), and the various forms of personal feedback that people receive from self-tracking devices (see (Lanzing 2018).
The transfer of nudging into the digital world may, at one level, be interpreted as it becoming less relevant to geographical concerns: as nudges migrate from the changing of actual environments to digital ones. But, the reverse appears to be true. The reason for this is that the emergence of hyper-nudging is directly coupled to the embedding of the digital into the everyday spaces of life. The emergence of smart devices (including mobile phones, smart, thermostats, the onboard computing capabilities of cars inter alia), couple with the embedding of a bewildering array of monitoring devices into the physical infrastructures of everyday life, is the very basis of the hyper-nudge. It is the Internet of Things (or the wideware) of the contemporary digital age that enables the data on which hyper-nudges depends to be gathered (perhaps relating to our particular penchant for running), and appropriate nudges to be delivered (perhaps a message to let us know of a particularly good deal on running shoes that is available in a shop you are in close proximity to). In a strange way, while the preliminary stages of the internet embodied clear demarcations between the realms of virtual and real space, it appears that in the future computing will be so deeply embedded in the world around us that it will be inherently geographical.
Of course, the fusion of the digital and the geographical has already been described in the pioneering work of certain digitally-oriented geographers (see Thrift and French, 2004). In their account of the automatic productions of space and the associated technological unconscious, Thrift and French consider how software is insinuating itself into our everyday lives and offers forms of local intelligence that is reshaping our worlds. Thrift and French draw particular attention to how the work of software in urban space tends to operate below the threshold of the representational, and the particular political and epistemological implications this presents to urban studies (Thrift and French (2002)) 312, Quoting Hansen, 2000: 17). Thrift and French interpret the spread of software into everyday life as the emergence of a form of distributed cognition in and through which our everyday environments become contexts for increasingly diverse forms of knowledge production and non-human analysis. While Thrift and French’s analysis is portentous of the age of surveillance capitalism we now find ourselves in (Zuboff 2019), it was written in a time before the proliferation of smart devices, cloud computing, and social media platforms. As such, while Thrift and French’s work identifies the spatial implications of governmentality by software, it did so in an age before the emergence of the hyper-nudge. To these ends, there is clearly scope to reconsider the ways in which data and the behavioural sciences are insinuating themselves into the spaces of everyday life, and the particular geographical implications of these processes. In what remains of this short reflection, I will consider the role of geography in the developing story of hyper-nudging. I will also consider how geography can provide us with a framework and language to develop novel critical perspectives on, and grounds for resisting, these emerging processes.
Unpacking the geographies of hyper-nudging
One does not have to look very hard to discern the emerging geographies of hyper-nudging. It is important, however, to identify three key processes that are associated with the spatial hyper-nudge. First are the processes I describe as spatial behavioural surveillance (catchy, I know). Spatial behavioural surveillance is different from more general forms of digital surveillance to the extent that it is not limited to the virtual world. Spatial behavioural surveillance involves the gathering of behavioural data in geographical situ. Spatial behavioural surveillance can operate on an aggregate (changing patterns of driving behaviour monitored through embedded kerbside devices) and more personal levels (as GPS activated smart watches indicates to Strava the particular routes we like to cycle to work along, or photographs are “geo-tagged” in Facebook). Spatial behavioural surveillance involves both locational behavioural surveillance (identifying what behaviours happen where) and the surveillance of geographical behavioural routines (identify the spatial routes and patterns of everyday life).
The surveillance of geographical behavioural routines: Strava Heatmap for my local area (including some of my own particular routes)
The second set of processes associated with digital hyper-nudging is geographical digitization. Unlike spatial behavioural surveillance this process is not primarily interested in behaviours in space, as enabling the geographical (in its entirety) to have a digital form. The transferal of the geographical into the digital is perhaps expressed most obviously through the Google Earth and Google Street View projects, in which satellite and ground-level surveillance facilitate the production of digitally adaptable maps of the world. Geographical digitization can operate at large and small scales. So, while Google Earth and Google Street View involve the orchestrated extraction of geographical data at large scales, Google Glass offered more personalised, and invasive, forms of surveillance, whereby anyone wearing Google glasses could help to map the world for Google in real time. But geographical digitization can take an even more sinister form. Zuboff (2019) describes the capacity of iRobot’s autonomous vacuum cleaner to develop detailed floor plan information of homes while cleaning rooms (235). It is claimed that there could be quite a market for household floor plan data in the future (although I must admit to finding it difficult to understand what the commercial benefits of this data would be – “answers on a postcard please”).
Whether generating the God’s-eye perspective of Google Earth, the more dynamic mappings of Google Glass, or iRobot’s more intimate floor plan data, geographical digitization provides the spatial co-ordinates in and through which hyper-nudging can be most effectively mobilised: in order to be able to hyper-nudge geographically, you first need to have a digital version of the geographical world. It is these digital coordinates that enable the activation of geographically salient knowledge, not only at the right time, but also in the right place!
Google Street View Camera in action (Tech Guide)
The third and final dimension of the geographies of hyper-nudging is the spatial hyper-nudge. The spatial hyper-nudge is essentially a non-coercive prompt to action that is based on algorithmically determined correlations of personal and collective data sets that are able to predict what you would like to do next and route your behaviour accordingly. Spatial hyper-nudges are distinct from more general forms of hyper-nudge to the extent that they are determined by spatial context (as opposed to the webpage that you are on), and target spatially salient behaviours. Spatial hyper-nudges could be used to maximise the behavioural possibilities of particular settings: for example, letting you know that a car park you are passing has available spaces, that a swimming pool is currently fairly quiet, or that a nearby restaurant is loved by one of your friends). But they could also be used to actively route you in certain directions. It has been suggested, for example, that the game Pokemon Go has been using play as a way of guiding people to particular, fee-paying, commercial establishments (such as McDonald’s) where Pokemon will be waiting (Zuboff, 2019). Google’s Sidewalk Labs is also promoting a new approach tackling traffic congestion in cities, which combines AI and smart tech to guide people to available public and private parking spaces (a kind of Airbnb for parking); this tech can also, however, guide traffic wardens to lucrative areas of cities (Guardian, 2016).
While a systematic analysis of the particular ways in which geography and nudging are, or could be, combined in spatial hyper-nudging is still to be complete, the possibilities are intriguing. We may only just be seeing the beginnings of the ways in which spatial hyper-nudging can deploy the behavioural tools of defaults (assumed best places), social influence (the spatial behaviours of those that you know), or status quo bias (you have been here before, so why not go again). What is clear is that smart tech, the Internet of Things, and machine learning, will enable nudging to enter the spaces of our everyday lives in ways it never has before.
Critical geographical perspectives on the hyper-nudge.
In this final section I want to offer some critical perspectives on hyper-nudging, which are specifically signalled by a geographical perspective. Much critical analysis of hyper-nudging focuses on the question of privacy (Yeung 2018). Concern over privacy emerges because of the large volumes of personal data that must be harvested to feed the algorithms that support hyper-nudges. Critically, however, Lanzing (2018) that the hyper-nudge involves a new horizon of privacy concerns as infringements on informational privacy (perhaps pertaining to demographic profiles) are joined by those pertaining to decisional privacy. Decisional privacy is distinct from informational privacy as it does not just tell us about a person (data which can be used to predict behaviour), it reveals the actual way a person behaviours in a given context. Access to decisional data (which can be both promoted and harvested by the smart devices that facilitate hyper-nudges) opens-up new realms of behavioural experiments and controlled trials at previously unattainable scales. The operation of such trials, often without meaningful forms of consent, raises a series of troubling ethical issues (Jones and Whitehead, 2018). Concerns over decisional privacy have two primary geographical dimensions. First, whether it be through the biometric reconnaissance of wearable tech, or the domestic surveillance of the smart home, hyper-nudges have enabled the behavioural governance of decisions to enter the everyday spaces of life in ways that analogue nudges never could. So, whether it be your smart fridge nudging you to consume less food, or your smart car encouraging you to drive more sensibly (and secure better insurance premiums), hyper-nudging changes the geographical scope of soft paternalism and the monitoring of related behaviours. Second, the GPS activated technologies that are associated with hyper-nudging do not only capture decisional actions in particular places, they can also monitor spatial action itself. Hyper-nudge technologies can compromise decisional privacy to the extent that they can monitor and mould our spatial decision-making and the routes we take (see above). While not directly related to hyper-nudging, one of the most striking recent violations of spatial behavioural privacy has been perpetrated by Uber. In its infamous ‘Rides of Glory’ analysis, Uber was able to map users one-nightstands (without their meaningful consent) on the basis of historic spatial patterns in customers use of the ride sharing app. What Uber’s ‘Rides of Glory’ analysis reveals is the ability of GPS enabled smart tech to not only reveal private spatial behaviours, but also to use patterns of spatial movement as proxies for the prediction of actual behaviours. This inferential capacity means that observed spatial behaviours can be used to reveal, and potentially mould, the behaviours that are occurring in between geographical movements.
A geographical perspective on hyper-nudging also highlights its exploitative potential. By being able to simultaneously cross-reference “geo-tagged” data concerning location, biometrics, and historical behavioural patterns, hyper-nudging has the potential to be able to exploit people’s spatial behavioural vulnerabilities in previously unimaginable ways. Being able to nudge you in the direction of a fast-food restaurant on the basis of the basis of time of day, proximity to a fee-paying eating establishment, biometric data on your level of hunger, and historical data on your penchant for hamburgers, could enable nudges to wield new forms of emotional power. This, of course, is the realm of mobile life-pattern marketing (see Zuboff, 2019: 242-245). As such it is perhaps best termed “hyper-sludging” (the use of nudge techniques for commercial gain, see above). Nonetheless, it is clear that the proliferation of ‘context aware data’ facilitates the exploitation of human spatial vulnerabilities in enhanced ways. While this may improve the effective of nudges, it is likely to significantly reduce our ability to be able to resist them. We can log off from a computer, but it is much more difficult to opt out of the internet of thing.
Biometrics and context aware data: Personalised, pop-up advertising in the Film Minority Report (2002, 20th Century Fox, screen capture)
So far, we have seen how a geographical perspective on hyper-nudging can help to provide critical vantage points on how big data and behavioural science are insinuating themselves within the spaces of everyday life. It is also clear, however, that geography offers some scope for collective resistance to hyper-nudging. According to Zuboff (2019), the growth of the surveillance practices associated with hyper-nudging have progressed through the geographical logic of trespass. The logic of trespass has seen the big data industry moving as far as it can into the personal spaces of existence until political or legal opposition is encountered (at which point such opposition is strenuously resisted). Whether it be in relation to the extraction and sale of data concerning online behaviour, or the photographing of private homes as part of Google’s Street View initiative, trespass has provide a lucrative model of economic accumulation (particularly when individuals are often unaware of acts of trespass and feel disempowered to do anything about it). The ethos of trespass is tightly connected to the Californian Ideology that undergirds the big data industry and suggests that only by doing things ‘without permission’ can the creative liberalism of the tech industry be realised (Barbrook and Cameron 1996). It appears, however, that when trespass takes on a physical geographical mode of action that opportunities for resistance are enhanced. In 2010 for example, residents of London Road, a cul-de-sac in Milton Keynes (UK) come together to prevent a Google Street View car from gaining access to the street. Local Councillor Edward Butler-Ellis explained the motivation behind the spatial protest on the following terms:
“The fact is they should have asked or at least let people know that they were photographing their houses. What really gets me is people have to opt out of being on it when they should have to opt in. A lot of older people without the internet are unaware that they are able to opt out of this.” (quoted in Barnett and Beaumont 2010).
Protests against Google in Kreuzberg Germany (Sean Gallup, Getty Images)
While London Road would eventually be captured on Street View, community action has seen many on similar streets and communities seeking to opt out of the forms of digital surveillance that provide the virtual infrastructure for hyper-nudging. The question remains, however, as to whether the notion of trespass can offer a spatial discourse in a through which more opposition can be generated to the embedding of digital surveillance and hyper-nudging into homes and communities. While there is no guarantee that resistance to the apparatuses of hyper-nudging will be more sustained in the real world as opposed to the digital world, it is clear that as digital surveillance moves from the virtual to the real that new opportunities for spatial resistance will emerge.
Barbrook, R. and Cameron, A. (1996). “The Californian Ideology.” Science and Culture 6(1): 44-72.
Barnett, E. and Beaumont, C. (2010) “Buckinghamshire village in Street View fight against Google” The Guardian March
Dow Schull, N. (2016). “Data for life: Wearable technology and the design of self-care.” BioSocieties: 1-17.
Jones, R. and Whitehead, M. (2018) “Politics done like science’: Critical perspectives on psychological governance and the experimental state” Society and Space 36: 313-330
Lanzing, M. (2018). ““Strongly Recommended” Revisting Decisonal Privacy to Judge Hypernudging in SElf-Tracking Technologies.” Philos. Technol.
Sunstein, S. R. (2019). On Freedom. Oxford.
Thrift, N. and. French, S. (2002). “The automatic production of space.” Transactions of the Institute of British Geographers 27: 309-335.
Yeung, K. (2016). “‘Hypernudge’: Big Data as a Nodue if Regulation by Design’ ” TLI Think! Paper 28/2016.
Yeung, K. (2018). “Five fears about mass predictive personalization in an age of surveillance capitalism ” International Data Privacy Law 8(1): 258-269.
Zuboff, S. (2019). The Age of Surveillance Capitalism: The Fight for a Human Future at the New Frontier of Power. London, Profile Books.
2 thoughts on “The Geographies of Hyper-Nudging: Navigation, Routing, and Trespass”