The deployment of digital infrastructures in cities increasingly mediates city life and people’s access to urban and public services. Underpinned by various entities, this mediation often adopts a hegemonic approach to data production, in which the entity controlling the service or infrastructure controls the produced data de facto (Carballa Smichowski, 2019). Given that data collection schemes in cities are often realised through the enclosure of digital infrastructures or Public-Private Partnerships (PPPs) (Barns et al., 2017; Morozov and Bria, 2018), the problem with a hegemonic model stems from conflicting interests between the private sector (that holds the power in a deregulated data ecosystem); the public sector; and people living, working, or studying in cities (on whom these initiatives might be imposed). Data-related policies and regulations targeted personal, identifiable data through reforms characterised by notice and choice regimes (exemplified by opt-in/opt-out options) and rights granted to data subjects, including access to data, rectification, and the right to be forgotten (Goldenfein and McGuigan, 2023). However, these measures are often considered inadequate as they fail to address the social and relational aspects of data from which value is derived (Viljoen, 2021; Zygmuntowski et al., 2021). This deficiency is further amplified as disruptive technologies, such as artificial intelligence, become more embedded in cities, posing challenges related to human autonomy and privacy and exacerbating existing social inequalities (Da Silva Carvalho et al., 2023; Calzada, 2018, 2023; Foth et al., 2021; Milchram et al., 2020). Therefore, there is an urgent need for governance approaches that anticipate the deployment of such technologies and address these associated risks (Micheli et al., 2020; Sanfilippo and Frischmann, 2023).

In this light, criticism has been raised regarding the absence or dilution of people’s involvement in shaping the conditions of data production and asymmetries in the distribution of relevant benefits (Artyushina, 2020; Cardullo and Kitchin, 2019a, 2019b; Kitchin and Lauriault, 2018). This has prompted scholars to rethink the role of people in relation to data production in the city, captured here in two key arguments. The first regards people as co-producers of data and argues for the acknowledgment of their contributions to the value created from it and their inclusion in setting the conditions of its production (Arrieta-Ibarra et al., 2018; Ducuing, 2024). The second, which is rooted in Critical Data Studies, highlights how data is not merely representative of the city but rather plays a role in (re-)producing it, indicating the necessity of including people in data-related decisions to fulfil the right to the city—the right to shape the city that shapes them and their lives in return (de Lange, 2019; Harvey, 2008; Kitchin and Lauriault, 2018). Against this backdrop, the UN-Habitat’s Flagship program, People-Centred Smart Cities, emphasised the need to put people at the core of digital transformations and corresponding data governance (DG) models (UN-Habitat, 2021), which determine power relations between entities impacted by or impacting data collection, control, sharing, and use (Micheli et al., 2020).

People-centredness broadly implies the incorporation of needs and perspectives of people into the development of systems. The UN-Habitat (2021) mentioned pillars of people-centredness such as inclusion, equity, empowerment, security, and participation. Nevertheless, its materialisation at the level of DG models in cities remains unclear. As evidenced by the literature, this ambiguity might be attributed to the absence of established terms in the field (Cardullo and Kitchin, 2019b; Liu, 2022; Micheli et al., 2020). Specifically, the confusion about what constitutes people-centred data governance (PCDG) might be stemming from the lack of explicit definitions, coupled with contestations over whether models suggested as PCDG, implicitly and explicitly, incorporate people’s needs and perspectives in meaningful ways (Cardullo and Kitchin, 2019a; Lehtiniemi, 2017; Micheli et al., 2020; Rinik, 2020). This work aims to address this gap by grounding the concept of PCDG in the city. It introduces an evolving framework that consolidates and elaborates people-centred notions at the level of DG. To achieve this, a critical scoping review of PCDG in cities is conducted. The review starts with preliminary indicators of PCDG, shaped by literature. Through an iterative process, it aims to understand PCDG on a nuanced level, highlight its pillar aspects, and reassess it. The crux of this exploration rests on the presupposition of DG as a socio-technical system. The paper aims to answer three questions: (1) What are the overarching aspects of PCDG in the city? (2) What are the archetypes of a PCDG model in the city? (3) How can PCDG in the city be improved?

The definition of DG adopted in this paper is borrowed from Micheli et al. (2020). DG is perceived as a socio-technical system that determines (1) the power relations between entities and individuals involved in or affected by data collection, control, sharing, and use; (2) the value to be derived from data; and (3) the distribution of benefits (Micheli et al., 2020). The study focuses on four moments of data flow: (1) the conception of data which includes collection and other processes that shape its formation (e.g., decisions on what is datafied and corresponding investment schemes); (2) control, which involves exerting control over data and access to it; (3) sharing, which encompasses sharing mechanisms, protections, and conditions; and (4) the realisation of emergent benefits, which presupposes data use but focuses on shaping and deriving benefits that emerge from the data apparatus as a whole. The remainder of the paper is structured as follows. First, the methodology of the review is outlined. Second, the extracted themes and archetypes are presented and discussed. Finally, the conclusion is presented.

The study started with a preliminary review of the literature, which informed the search strategy and data extraction framework. This section presents these along with the methods corresponding to coding and analysis.

PCDG is a multifaceted concept that can materialise in various ways. The findings indicate that PCDG in the literature is attributed to one or more of the following four characteristics: (1) a set of values that are either inherently people-centred, such as inclusion and privacy, or that contribute to the people-centredness of the process or goals, such as openness and transparency; (2) the inclusion of people as agents, beneficiaries, or enablers of DG; (3) the incorporation of mechanisms for people engagement in DG; or (4) the alignment of the model’s claimed goals with people-centredness. Table 2 highlights the three most common themes in each studied category, providing a glimpse into potential PCDG aspects and mechanisms. The 6 archetypes shown in Figure 4 are constructs extracted from the collective of reviewed passages aiming to conceptualise representations of DG. They are not intended to inform discrete implementations of DG models in cities: multiple constructs may have a nested relationship or complement each other (e.g., Calzati and van Loenen, 2023a; City of Barcelona, 2015; Fischli, 2022), or some models could incorporate elements of these archetypes rather than adopting them in their entirety (e.g., Foth et al., 2021). The results of this study do not imply completeness but are part of an evolving PCDG framework that should be complemented by further aspects and constructs. Nevertheless, they contribute to discussions about desirable urban futures by unpacking the notion of PCDG as it exists in the literature, providing a foundation upon which interpretations can be problematised and normative work can be based. The aim of this work includes a reassessment of PCDG. To achieve this, the remainder of this section critically examines the archetypes, identifying both problematic and constructive components. Finally, insights from this critique are consolidated to inform a reassessment of PCDG.

Both CompA and RBA represent models that adopt individual-centric approaches, focusing primarily on the data subject. The first pitfall of such models is their exclusion of mechanisms pertaining to the conception of data. They focus on how data can be controlled, owned, or sold by the data subject rather than why and how it is collected (Doned and Belli, 2020; Lehtiniemi, 2017). Addressing the conception of data requires collective mechanisms rooted in democracy and civic participation, which go beyond an individual’s capacity to regulate their own data flows. Instead, they involve collective decisions about which issues should be datafied (de Lange, 2019) and investments in data collection schemes (Muldoon, 2022). Indeed, in many cases pertaining to CompA, the data subject decides whether or not to collect data (e.g., Wang et al., 2014). However, the scope of this decision-making is typically limited to a simple “yes” or “no,” as the specifics of what to datafy and expected benefits are usually predetermined by potential data consumers.

Furthermore, emergent benefits are realised only when data exits the associated models’ domains. For example, while CompA aims to promote economic fairness by allowing the data subject to intervene in the extractivist data regime (Lehtiniemi and Haapoja, 2020), a fundamental issue undermines the viability of compensation as a people-centred solution: compensating for data turns it into a private asset, primarily serving the interests and goals of the entity that comes to own it and separating people from the realisation of the emergent benefits of data. In RBA, the motivation for data sharing must be carefully considered if the emerging benefits are to be realised by data subjects and the general public. Altruistic endeavours or self-interest, coupled with the capacity to engage in the activity of sharing, often drive data’s flow outside the model’s domain to realise these benefits. The latter requires not only an intrinsic willingness but also the presence of external initiatives that align with the values and goals of the data subjects.

In CivRA, data subjects are represented by an intermediary, which is tasked with governing initiatives by both the private and public sectors and stewarding the ‘public interest’. Although models associated with this archetype are an improvement to the status quo, they might not ensure the incorporation of people’s perspectives and needs, nor guarantee that benefits are realised by them, for three main reasons. First, the archetype is recognised as the most conservative amongst the 6 archetypes in terms of people’s participation and control. It functions by representing citizens and their interests without their direct involvement, contrasting sharply with others that facilitate data-related decision-making by data subjects or collective deliberation. The preference for representation in CivRA over deliberation is based on the assumption that deliberative processes are impractical (König, 2021; Ryfe, 2005). However, this approach to democracy often aims to preserve the core of the system rather than challenge it (Blaug, 2002). In contrast, deliberative processes are viewed as more legitimate in integrating people’s needs and insights into decision-making (Blaug, 2002; de Hoop et al., 2022), thereby fostering an environment in which the emergence of ideas that could potentially challenge existing systems is possible.

Second, weak participation coupled with an arrangement of agents including the public and private sectors along with citizen representatives might reproduce a microcosm of the current governance landscape. This arrangement, especially when (1) the trust itself is developed and enabled by private entities, (2) initiatives involve Big Tech, or (3) initiatives are underpinned by Public-Private Partnerships, might fail to balance power dynamics or reduce asymmetries (e.g., Artyushina, 2020; Sidewalk Labs, 2018). These dynamics might be further reinforced by the vagueness and problematic application of the term ‘public interest’. For example, Short (2023) highlighted that the ‘common good’ or ‘community’ values are seldom considered by American agencies in analysing the ‘public interest’. Instead, an outcome is often considered to align with the ‘public interest’ if it is justified by “studied, economic arguments” (e.g., cost–benefit analysis) (Short, 2023, p. 759). Moreover, the prevailing interpretation of ‘public interest’ tends to adopt a majoritarian view, which can leave minority and marginalised communities unprotected (Feasby, 2020). This is especially problematic since vulnerable groups are already disproportionately impacted by the datafication of the city (Kennedy et al., 2021; Tracey and Garcia, 2024).

As seen in ResCA, DDA, and CommA (Table 3), it is expected that the domains of models associated with data as a common good encompass the realisation of emergent benefits. This concept extends beyond segmenting access of data between a group and outsiders, materialising in the realisation of data “for the common good” (City of Barcelona, 2015, p. 27) and the eventual “production of a common good” (de Lange, 2019).

ResCA positions the public sector as a defender of people’s right to the city, working to balance power asymmetries between corporations and citizens. While a traditional concept of openness, operationalised through open data initiatives, likely exists in cities adopting such a model, ResCA promotes a form of openness that demands action from private entities. This is enforced by the public sector through mechanisms like procurement agreements, public tenders, or licensing terms that stipulate interoperability and mandate the sharing of citizens’ data (City of Barcelona, 2015). The aim of ResCA is to transfer control of data and digital infrastructures back to the people and to the city itself (City of Barcelona, 2015; City of Amsterdam, 2021). It does not accommodate traditional corporate practices, particularly because it aims to counteract these forces.

DDA benefits from strategies and policies in cities that support alternative economic actors and social organisations. In return, it enriches the city by promoting fairer labour environments (DDA.A) or offering personalised options for managing personal data streams that leverage contributions of data subjects for the greater good (DDA.B). However, this model is viewed as complementary rather than standalone due to several limitations: (1) it lacks mechanisms for the initial conception of data; (2) it relies on public sector support to counteract competitive pressures from corporations and to encourage data donations from individuals; (3) although it has the potential to contribute to systemic change and advance the common good, it primarily serves the immediate interests of co-op members or data donors; and (4) it does not address other forms of urban data that are not identifiable.

CommA is driven by people from the conception of data through to the realisation of its emergent benefits. It presupposes the development of a shared understanding of what constitutes the common good and how it can be achieved through the collection and use of data (de Lange, 2019). Although this model serves as a form of resistance to the status quo by enabling the community to meet its own needs and shape its narrative, it often offers an alternative to corporate practices without actively disrupting them on a larger scale. Effective forms of societal change at the city-level require both: building alternatives and abating existing hegemonic practices (Hebinck et al., 2022).

The three aforementioned models (ResCA, DDA, CommA) realise data as a common good through different configurations of their Normative, Actors, and Technical layers. However, they share three key commonalities. First, the power dynamics underpinning each of these socio-technical systems favour people: the public sector positions itself as a defender of the people against the hegemonic model by focusing on public-community partnerships (ResCA); intermediaries act solely as fiduciaries to ensure the benefit of those who co-produced the data (DDA) or those impacted by it (DDA.A); or people serve as enablers, agents, and beneficiaries within the DG system (CommA). Second, these models play a role in balancing power asymmetries imposed on cities by the hegemonic ownership model by (1) countering practices that inhibit people’s control over datafication (ResCA) and (2) developing alternatives to the status quo (ResCA, DDA, and CommA). Third, radical forms of people’s engagement in DG are demonstrated in these models, including co-creation and deliberation. These forms connect people directly to decision-making spaces around key aspects of the system such as its ontology (e.g., defining what the common good entails), goals, and beneficiaries.

This study unpacked and reassessed PCDG in cities through a systematic scoping review. It extended the discussion of PCDG beyond initially identified indicators, such as participation and instrumental aspects. It emphasised the importance of connecting people to the emergent benefits of data, the power dynamics within the model, the extent to which participation facilitates decision-making, and the model’s role in balancing power asymmetries within the broader landscape. The conclusions of the paper thus argue for (1) moving away from compensation models, which separate people from the emergent benefits of data and legitimise its use as a private asset; (2) exercising caution around rights-based models as standalone solutions, suggesting their integration within broader initiatives that treat data as a common good and encompass both the conception of data and the realisation of its benefits; and (3) a shift in focus from pursuing a vaguely defined and often majoritarian ‘public interest’ to embracing the ‘common good’, which is defined by the people and provides a consistent normative framework throughout the DG model.