In the following five subsections, we highlight our key observations with regard to each of these five dimensions, discuss their implications for smart knowledge politics and raise suggestions for the realization of more inclusive deliberative spaces for smart knowledge politics.
...engaging with politics of inclusion: who
Some smart projects are more inclusive than others in terms of who gets to participate. We observed discrepancies between whose voice is considered legitimate and who is affected as well as differences to which extent diverse knowledges were taken seriously. These discrepancies build on and emerge from existing power relationships. Indeed, we also observed how the timing, location, use of language and jargon, group composition, skills required to understand and contribute to deliberations as well as institutional boundaries and responsibilities produced inclusions and exclusions of actors and what they can(not) say.
Discrepancies between whose voice is considered legitimate and who is affected were observed frequently. For example, Hamburg’s city administration effectively, even if unintentionally, excluded publics and their differently experienced concerns about mobility policy from their activities to realize smarter mobility planning by substituting users’ voices with observational data on their movements (Späth and Knieling 2020). In terms of Cardullo and Kitchin’s (2019) ‘scaffold of citizen participation’, this implied that citizens did not participate through consultation or a partnership (had citizens’ voices been considered legitimate) but as consumers through choice or even as non-participants through the use of their data about the way they made use of the city’s mobility system. In response, we propose that whose voice is considered legitimate, in what form and on what basis needs to be rendered explicit and subject to debate.
Such inclusions and exclusions emerged from prevailing power relations and embedded in smart city project, as well as from smart projects’ organizational structures, operational processes, ownership rights and so forth, and from the digital affordances of the technologies deployed (also see e.g. Cowley and Caprotti (2019) on the role of power relationships and e.g. Calzada and Cobo (2015) on the role of digital technologies). Who initiates and coordinates smart city projects, and these actors’ relationships with other stakeholders, both play an important role in who gets to say what about the issue and its framing.
For example, discussions with citizens on what smart data would be relevant did feature in projects initiated by civil society, such as the GammaSense project in the Netherlands and the Making Sense Project in Barcelona, while dusch discussions did not feature in the EU-funded and municipality-initiated and coordinated project MySmartLife in Hamburg (de Hoop et al. 2022; Späth and Knieling 2019). Notably, different stakeholders perceive these power relationships differently. For example, where municipal and company representatives in the Dutch Jouw Licht op 040 (‘Your lighting at area 040’) project in Eindhoven felt they were providing citizens a much larger voice than usual, some citizens in the very same project considered their concerns to be systematically sidelined. Other examples illustrate how the production of information through smart technologies was observed to reinforce (e.g. MySmartLife in Hamburg; Späth and Knieling 2019) or subtly shift power relationships (e.g. GammaSense in the Netherlands; de Hoop 2020) by making some forms of data generation more actionable than others. Hence, we propose to be more attentive towards the role of power relations to regarding who facilitates and/or controls deliberation in a smart city project, the different kinds of data, information and knowledge that are considered and framed through this deliberation, and to what extent such deliberation explores how power relations may need to be transformed in order to accommodate for other relevant knowledges.
Inclusions and exclusions of actors and what they can(not) say further resulted from the timing, location, use of language and jargon, group composition, and from the skills required to join, understand and contribute to deliberations. Inclusive multi-stakeholder engagement requires resources, time, and the motivation to persevere, especially when difficulties or unexpected events arise (Evans et al. 2019). For example, in the Slim City project in Utrecht (de Hoop et al. 2019), outcomes were expected within a predefined time-frame of 10 weeks, which led to conflicts between actors who wished to speed up the process, typically those that had to report back to their organizations, and actors who felt the need for more in-depth deliberation. Hence, we propose that an active effort should be made on the part of major players such as government authorities, company representatives etc., to include people and perspectives that are currently being excluded. This may require sufficient availability of resources and skills among hosts and participants, appropriate timing and location of events, use of accessible and sufficiently open-ended deliberative techniques, training, careful composition of groups, and hosts may need to learn to listen to diverse publics expressing themselves differently, including in forms involving embodied and tacit knowledge, from what hosts seeking codified evidence may be used to.
Finally, in multiple projects we encountered that institutional boundaries and responsibilities, too, can exclude relevant actors from deliberations, while successful transgression of such boundaries may indeed result in the inclusion of a wider range of relevant actors (also see Nicholds et al. 2017). Departments, sections, divisions etc. delineate responsibilities and often stipulate who is central and who is peripheral to the project – even when the boundaries between different parts of an organization do not reflect the multi-dimensional and cross-domain nature of challenges addressed by smart urban initiatives. Hence, we propose the need to reflexively navigate and reconfigure institutional boundaries and responsibilities in order to include relevant actors from all relevant domains.
...engaging with politics of recognition: what
The focus of smart knowledge politics deliberations - the what - entails setting boundaries, attentiveness towards the processes that shape what can and cannot be said and reflexivity regarding the values and priorities embedded in knowledge production. First, we observed how the use of varying, changing and sometimes clashing definitions and operationalizations of key terms had implications for the way in which knowledge could (not) be produced. Second, the empirical focus of a particular project with smart urbanism interacted with the ‘who’ discussed in the previous section. Third, in many projects integral deliberation was challenging in the context of smart urbanism experiments’ effects cutting across institutional boundaries and communities of practice. Non-smart knowledges and solutions to identified problems as well as the strategies required along with data production to realize ambitions were often overlooked. Finally, a persistent belief in the objectivity and neutrality of facts obstructed discussion on the values and priorities embedded in smart knowledge production.
First, with regard to the diverse, clashing and evolving definitions and operationalizations of key terms (such as inclusivity, democracy, sustainability, privacy, human rights, safety etc.), we observed that which and whose definitions of these terms dominated had profound implications for the production of knowledge and the governance (in)actions that are produced, and whose interests these knowledges and actions serve (see also Echebarria et al.’s (2021) recent literature review on this point with regard to the way smart cities themselves are defined). In a different vein from Echebarria et al., who suggest working towards a single, albeit comprehensive, definition of the smart city, we propose explicitly discussing and jointly defining the meaning and the implications thereof in a way that is appreciative and respectful to systematic differences in positions towards these terms (cf ad hoc uses of buzzwords). This resonates with Cardullo’s (2021) observation that if citizen participation in smart urbanism is to be beneficial to these citizens, rather than to corporations developing smart tech, the purposes for which technological solutions are being developed need to be opened up for deliberation and meet citizens’ needs.
Second, the empirical focus on experiments with smart urbanism - the ‘what’ in terms of content - played a role in the way in which, and the extent to which diverse stakeholders deliberated the knowledge politics of such projects. t Whether this topic wasa direct concern to stakeholders involved, and the extent to which stakeholders felt this topic was their responsibility were particularly important here (also see, e.g. Ehnert et al. 2022). Very few Dutch citizens deliberated the development of the GammaSense tool to measure gamma radiation in the Netherlands, because at the time, they were not concerned with the risk of being exposed to gamma radiation (de Hoop 2020). In contrast, in the city of Utrecht, citizens concerned with the development of a high-density smart neighborhood near their homes unsolicitedly and loudly engaged with the knowledge politics around the city’s calculations on the project (de Hoop et al. 2019). Furthermore, what aspect is put up for deliberation - ranging from deliberating what a smart technology collects data on, how, and for what purpose, to deliberating only one of these aspects or choosing between a small set of pre-selected options (de Hoop 2020) - also played a critical role in shaping the associated knowledge politics deliberations.
Third, as observed in Section ”...engaging with politics of inclusion: who”, issues of concern that relate to experimenting with smart urbanism tend to cut across existing institutional boundaries and communities of practice with different – sometimes clashing, sometimes complementary, often interdependent – approaches towards the issue at hand and its potential solutions. Although we were not able to identify substantial literature on this issue, we for example observed how the development of a multi-purpose sensor-network around a busy traffic junction in the city of Eindhoven required different departments of the city council and partnering companies to collaborate in jointly designing the sensor-network. This, in turn, required that actors realized how their input into the design of the system enabled and foreclosed the possibilities of others who were responsible for different aspects of the same system. This was a highly complex and unusual endeavor for the actors involved, and supporting infrastructures to do so were absent. We therefore propose the need for knowledge politics deliberations to explicate, engage with and build stronger interrelationships between different departments, areas of expertise, and plurality of knowledges about the phenomena that is to be governed in a smart way.
More fundamentally, we observed that in smart urban experiments, the deliberations about knowledge politics that took place as well as (critical) scholarship on smart urbanism rarely engage with other (non-smart, non-digital) knowledges, as detailed more elaborately by de Hoop et al. 2022. However, the production, circulation and use of digital and non-digital knowledges can take quite different forms, and each will frame and approach the urban challenge differently. Various forms of knowledge may also work together, and therefore need to be deliberated jointly.
For example, a community-led citizen science project that used open-hardware sensors to monitor noise in a public square in Barcelona found that enrolling participation and producing the data required deep sociological knowledge about life in the square. And whilst residents as ‘smart citizens’ felt empowered by the noise data they gathered, subsequently acting on that evidence required political knowledge to mobilizing pressure for change and architectural and social knowledge about actions that could curb the noise nuisance, including knowledge about the leisure economy implicated in changes in the neighborhood associated with the noise data (de Hoop et al. 2022). Gabrys (2020) speaks of ‘creaturing data’, or the way in which data become creatures through perceiving and participating in environments in specific, non-digital ways. In addition, we observed that project ambitions may actually be reached in a more desirable, effective manner through non-smart means – something which is only possible when deliberation drives the selection and design of technologies after better specifying the ambition (also see Hollands 2015). This happened in the case of designing digital platforms for direct democracy in Madrid and Barcelona; Smith and Martín 2021). The success of the platform relied upon complementary offline processes, including old-fashioned community participation spaces and activities that mattered most to citizens. We therefore propose decentering smart technology and placing the ambition of the project in the lead instead, in deliberations which recognize and include a plurality of knowledges that may potentially be complementary to or more appropriate than smart knowledges, as well as both smart and non-smart pathways to reach participants’ ambitions.
The successful production of data was often equated with a successful project as a whole – even if the project may be discontinued because the data was of little use-value for stakeholders, such as the Fietstelweek (‘Bike counting week’) in Utrecht (van Oers et al. 2020). In such situations, one risks losing sight of the original project ambitions, as the allure of generating large data sets becomes an end in itself, disconnected from achieving urban change. We therefore propose making explicit what is needed to realize the changes aimed for beyond solely generating data.
Finally, all projects teach us that deliberating knowledge politics is challenging in the context of a persistent belief in the objectivity and neutrality of (smart) knowledge. There is no evading the fact that agenda-setting, defining research goals, developing methods, analyzing results and rendering knowledge actionable all require making value-laden, and hence political, choices. We propose that it is important to recognize that knowledge is as much created as it is discovered, meaning that facts and values are intertwined throughout the process of producing and using knowledge.
...engaging with politics of space and place: where
When we write metaphorically about spaces for deliberation, it is not by chance that we activate a spatial vocabulary. We observed that the specific places in which smart city projects were conceived, planned and operated played a key role in how data was generated and what data emerged, and that this had implications for the ways in which smart city projects may be transferred or scaled to other localities.
While science students learn that the results of ‘objective’ measurements and respective conclusions differ depending on where you measure or how you delineate a sample area, this insight is often overlooked in debates about smart ways of knowing the city – except for those rare projects that aim at the creation of a ‘science literate’ public, like the GammaSense project (de Hoop 2020). Indeed, locality and boundary drawing are not only crucial for the quality of data, but also to understand how the locality shapes the data generation process and leaves traces in the data (Tironi and Criado 2015; Mörtenböck and Mooshammer 2020). At the same time, the creation of large datasets erases the visibility of these local particularities in the data, which in turn has political implications in terms of a shift in or reduction of the intended actionability of the data. For example, in Utrecht, the collection and aggregation of data on cycling flows within and around the city disconnected data points from the cyclists and their motivations to prefer certain routes over others. Instead, these aggregated data points were primarily understood as indications for potential infrastructural flaws. As a result, the data turned out to be considered ill-suited to help improve infrastructural decisions, which was the original aim of the project, and the project was halted (van Oers et al. 2020).
Indeed, data transfer and up-scaling information, which is only possible when local specificities are erased, is a dominant ambition across many of the projects studied, and such ambitions are likely to meet with pluriform locally situated socio-material obduracies and resistances (Lombardi and Vanolo 2015). Broader architectures of knowledge generation and decision-making are furthermore highly path-dependent and historically shaped, and therefore differ from place to place, reflecting local political practices, expectations and traditions. In sum, we suggest that all four dimensions of socio-spatial relations – place, scale, territory and networks (Jessop et al. 2008) - need to be systematically considered when deliberating the generation, aggregation, transfer and use of data. Taking these spatial dimension seriously requires attentiveness towards the other four dimensions of this paper as well: shifts in place, scale, territory and networks in a smart urbanism experiment (where) may also foster different kinds of inclusion (who), a different focus (what), different timing and timescales (when) and different processes of institutionalization (how).
...engaging with politics of time: when
The temporal dimension, i.e. the timing and the order of smart urban projects’ activities is important to consider. We observed that knowledge politics are evident in all stages of such projects, and that choices with regard to the timing of deliberation have effects on what the deliberation may or may not contribute towards. We also noted that the time-demarcated nature of most projects limited possibilities to deal with urban challenges in a comprehensive and long-term manner.
First, with regard to the role of knowledge politics across all stages of such projects, we observed that choices with important future implications for the potential effects on the socio-material urban fabric of a smart experiment are continuously made, for example with regard to problem formulations, methods of data collection and analysis, reporting and communication, critique and validation, translation between settings and circulation, and applications in institutional practice and decision-making (de Hoop 2020; also see Chilvers and Kearnes 2016). However, in the projects that we observed, open deliberation on these issues only happened occasionally, and often in the form of a specific event, like a kick-off brainstorm meeting with multiple stakeholders, as we have seen in the Utrecht Slimcity project (de Hoop et al. 2019), or with regard to specific aspects of the project’s and technology’s design, as we have seen in the GammaSense project in the Netherlands (de Hoop 2020). Instead, we argue that deliberation should be a continuous endeavor and on the agenda of multiple moments of interaction before, during and after the project.
Second, we observed that the timing of deliberation in the context of the specific temporalities of smart urban experiments has implications for what such deliberations may and may not contribute towards. We saw that early deliberation allows stakeholders to join in steering the dimensions and underlying assumptions behind a project and to critically discuss potentially irreversible effects of an experiment from the start, but that it may be difficult to identify relevant stakeholders as well as areas for future conflicts in these early stages. Especially in district planning, such as in the SlimCity project, future residents are difficult to target. And if they were included, like in Brainport Smart District in Helmond, it was difficult for them to oversee repercussions or contribute alternative options (de Hoop et al. 2019). Planning deliberations later, however, makes it difficult to still adapt the project to issues voiced during such deliberations. We suggest that the timing and frequency of deliberations should be carefully considered in the context of what role deliberation is envisaged to play in the overall project.
Finally, smart urban initiatives often take the form of experiments and time-demarcated projects, resulting in a way of dealing with urban challenges that hardly stretches beyond the timespan of the project, limiting the depth of knowledge politics deliberations as well. In addition, projects’ short timespans also had consequences for the timeframe that was up for deliberation as foresight was highly limited. This, in turn, had implications for the other four dimensions discussed, particularly with regards to the breadth and depth of the empirical focus of deliberations (what), whose concerns are deemed relevant (who), which locations and scales are taken into account (where), and for the institutionalization of such deliberations (how). Furthermore, projects limited timespans, and the limited timeframes that were up for deliberations, made it difficult to continue activities at the location of the project itself, let alone scale or translate the project to other locations. We therefore propose it is crucial to be aware of the project boundaries and to work on relevance and continuity of the experiment beyond its formal duration and scope. Urban problems often require continuous effort and deliberation instead of specific events or a series of scattered projects. Ongoing controversies and conflicts can be a fertile ground for deepening deliberation (Verloo 2018).
...engaging with politics of institutionalization: how
Overall, we found limited examples of explicit deliberation of smart knowledge politics in our studied cases, and spaces for such deliberation were rarely institutionalized. If such spaces were institutionalized, we observed that existing arrangements, such as institutional boundaries or the timing of deliberations (see “...engaging with politics of time: when”), played a key role in what knowledge politics can and cannot be deliberated upon (see “...engaging with politics of recognition: what”) and by whom (see “...engaging with politics of inclusion: who”). Finally, we observed that traditional technology assessment approaches may not be suitable to take the data produced through citizen science projects seriously as well as adequately address concerns about future forms of urban governance, ownership of knowledge and urban lifeworlds that may emerge from widespread application of smart technologies.
With regard to the limited instances in which knowledge politics were deliberated, we observed that ways in which these deliberations took place differed substantially, ranging from formal approaches though public-private partnership negotiations (Smart District Eindhoven) or policy programs (Smart Hamburg) to emergent or uninvited deliberations such as in citizen science projects (GammaSense) or citizen protests (SideWalks project). Crucially, such plurality of formats allowed for different voices to be heard and issues to be discussed. Hence, we suggest that a more explicit institutionalization of knowledge politics deliberation can be helpful, but inevitably emerges with and through the wider urban politics in which such deliberations are situated, which in turn require reflection. In particular, pleas for institutionalization may overlook the importance of plurality in deliberation, seeking to standardize and demarcate possibilities for deliberation instead. However, we argue that ensuring plurality as well as adaptability and appropriateness should be a core consideration when designing institutions for deliberation of knowledge politics. Such a plural approach was for instance visible in our Barcelona case studies, where strong ideas about democracy, citizenship and open-source and commons-based technology development have been productively integrated into the evolution of smart city platform policies and infrastructures.
The second observation is that existing arrangements in and through which spaces for deliberation are constructed and institutionalized also play a role in when, on what and by whom knowledge politics can be deliberated upon. On the one hand, we argue that such existing arrangements may need to be redrawn to enlarge the space for knowledge politics deliberations. For instance, Mysmartlife in Hamburg demonstrated that formal requirements from European funding programs excluded the possibility of early-stage deliberations at the regional and local level (Späth and Knieling 2020). The proposed implication is that major funding bodies should consider a more plural and distributed approach to engaging with different types of stakeholders as part of their funding requirements. At a more fundamental level, there is a need to more broadly reconsider the role of the state in smart knowledge politics within a plural social landscape. At the same time, the design of deliberative spaces will often need to be considered pragmatically in the short term, within the possibilities, constraints and transformations of existing institutional arrangements. A number of our case studies evolved in relation to European discourse, policies and funding arrangements towards the smart city, such as the Mysmartlife case in Hamburg discussed in this paragraph, which are likely to be an exogenous environment that cannot be reshaped at will for a specific smart project seeking funding.
Finally, we observed that existing approaches to what constitutes ‘good’ or ‘valid’ data were often ill-suited to provide room for a diversity of knowledges and data. For example, our case studies on nuclear radiation (GammaSense) in the Netherlands and on noise levels (Making Sense) in Spain demonstrated that data produced in citizen science projects often lack legitimacy within established institutional arrangements (de Hoop et al. 2022). Furthermore, existing technology assessment often focuses on specific risks and is therefore narrowly framed, which provides insufficient space to explore concerns over the forms of future governance, ownership of knowledge, and urban lifeworlds that widespread use of smart technology might enable. We therefore argue that ensuring plurality in the institutionalization for deliberating smart knowledge politics could also be shaped through strengthening technology assessment capabilities at various governance levels. We propose moving away from dominance by private actors and formal experts in the articulation of visions, needs and demands for smart knowledge tools, techniques and platforms to include a wide variety of stakeholders (Smith and Martín 2021). As a consequence, constructive technology assessment capabilities should consider a broader approach and include assessments made, formally and informally, in other realms of society, including in the forms of citizen science projects or citizen protests. We propose technology assessment practices should consider such forms of smart knowledge and advice procedures more explicitly, even if they may not live up to the dominant framings around quality standards for evidence.