Actually existing smart forests: A proposal for pluralizing eco-technical worlds

Abstract

The “actually existing smart city” is a familiar figure and concept within smart environments research. Usually, “actually existing” analyses challenge the often-hyperbolic proposals for smart cities by grounding research in built urban conditions. Cities are not exemplars of digital interconnectivity and seamless functioning, many smart cities researchers note. Instead, they are complicated sites of uneven urban development. Such research places speculative urban visions in contrast with empirical conditions to reflect upon and re-evaluate smart city proposals as they hit the ground. Building on this critical research, this article considers how smart environments are proliferating beyond cities to other milieus, including forests. By first reviewing “actually existing” smart cities literature along with “actually existing” references in social and political theory, we consider how to update and advance the “actually existing” analytic by revisiting the perceived rift between speculative and actual environments. Drawing on interviews with smart forest stakeholders, the second part of the article develops three examples of “actually existing” smart forests that demonstrate how inseparable the envisioning, making, and sustaining of smart environments can be. Even more than empirical conditions counteracting fantastical visions, we propose the “actually existing” analytic can be updated and mobilized to examine how plurality, contestation, and democratic participation are at stake not just in the lived conditions of smart environments but also in the abstract and provisional contours of computational technologies as they shape and transform milieus. We suggest that engagement with the plurality of speculations, material conditions, and practices is crucial for ensuring “actually existing” eco-technical worlds and relations are attuned to equity and environmental flourishing.

In environments around the world, digital technologies are remaking forests. In what some have dubbed “forestry 4.0,” in line with other industry 4.0 applications, digital technology, supply chain logistics, environments, and industry are meant to connect for more seamless functioning. One video from FP Innovation Canada captures this overarching approach to remaking forests. Digital technologies range from remote-sensing satellites to lidar generated imagery and inventories of forests to enable resource management and forecasting. Internet of Forest technologies produce real-time and connected forests optimized through big data that facilitates monitoring, harvesting, and replanting. Advanced procurement systems use multiple sensors, from GPS to machine sensors, along with automation, robotization, augmented reality, remote operations, intelligent transport, “unmanned” trucks and road sensors to keep forest products moving. Data analytics predict and respond to “client’s needs” along with decision support systems to ensure the “right wood” goes to the “right mill” (FPInnovations, 2017). While this forestry 4.0 vision is, in many ways, a speculative diagram of multiple smart forest technologies and capabilities, in other ways it describes technologies already in use. This video proposes and describes ways of connecting forests with digital technology, personnel, machinery, resource requirements, and industry objectives. Emerging smart environments such as forests are at once proposed, in the making, and already implemented. Practices for realizing smart forests unfold across these co-constitutive registers.

“Forestry 4.0” video. FPInnovations. 2017. https://www.youtube.com/watch?v=r4vhLQ8OEP0.

Most research on smart environments has focused on cities as the primary spaces where digital technologies have been proposed and implemented. Urban-computational infrastructures create “urban operating systems” (Luque-Ayala and Marvin, 2015). In addition to cities, many other smart environment proposals and projects are in varying stages of development, including prototype, beta, speculative or fully installed formats (Machen and Nost, 2021; Turnbull et al., 2023). Oceans drift with sensor buoys and autonomous submersibles (Helmreich, 2009; Lehman, 2018). Agricultural fields hum with robots and drones (Bronson and Knezevic, 2016). Forests interconnect through satellites, sensors, drones, and cameras that monitor forest health, extent, and change (Gabrys, 2020). Across these different milieus, digital technologies are remaking planetary environments through processes of automation and optimization, efficiency, and networked communication, thereby promising to become “smart” and even a “system of systems” (Gabrys, 2016). In other words, smart city technologies are materializing in numerous other environments.

While an impressive and incisive range of smart cities research has documented and analyzed the changing configurations of environments, technology, and social-political consequences, there are less extensive studies into smart environments such as forests. Forests are increasingly instrumented and joined up with digital technology, and they are also speculative projects in the form of forestry 4.0 that would automate responsiveness to climate and environmental change. Building on insights from aligned smart environments research, this article draws on established smart cities discourse as a stepping-off point for this study into smart forests. Within a vast range of smart city literatures, we zero in on one specific area that engages with the “actually existing smart city” (Shelton et al., 2015) to consider how smart forests materialize. As a familiar figure and concept within smart environments research, “actually existing” refers to attempts to challenge the often-hyperbolic proposals for smart cities by grounding them in built urban conditions (Shelton et al., 2015; Shelton and Lodato, 2019). Rather than the fantastical scenarios of a 170 kilometer-long “Line” smart city in Saudi Arabia or the rejected Sidewalk Labs in Toronto (Flynn and Valverde, 2019), the “actually existing” analytic engages with the complicated and uneven conditions of urban development projects. Such research places spectacular urban visions in contrast with built conditions to reflect upon and re-evaluate smart city proposals as they hit the ground, whether by analyzing how digital testbeds in post-industrial cities can reinforce inequalities (Dalton et al., 2020; Datta, 2018), or by examining how data dashboards can amplify surveillance (Barns, 2018; De Lange, 2018). In these analyses, cities are less a terra nullius and more a thicket of infrastructures, organizations, publics, resources, and political struggles (Levenda et al., 2020; Masucci et al., 2020; Tavmen, 2020). In this way, research that examines “actually existing” smart cities pivots from a study of plans and proposals to engage with empirical and lived smart city projects.

Building on, reworking, and extending the “actually existing” analytic, we investigate how “actually existing” smart forests take shape while revisiting the theoretical and methodological distinction that the “actually existing” orientation offers. By investigating how smart environments are proliferating beyond cities to other milieus, this article asks: How can other smart environments such as forests be examined through an “actually existing” smart cities analytic? How do “actually existing” smart forests materialize, whether in plans or projects? Finally, how do smart forests in the making rework the co-constitutive practices of speculative, lived, and “actually existing” environments? These questions seek to extend “actually existing” analyses to grapple with the speculative and material conditions of smart forests as they are built and coming into being. The perceived rift between speculative and built smart forests is a central focus in the pursuit of this inquiry. To this end, we revisit the focus on empirics and groundedness as critical features of the “actually existing.” In complement and contrast to current research, we consider how this conceptual frame returns to a re-engagement with smart forest speculations, as much as empirical conditions. Smart environment projects exist in a range of states of implementation that hover across and blur the limits of the speculative and grounded.

“Smart” is an ambiguous term—sometimes deliberately so (Hollands, 2008). Smart technologies increasingly convert environments into sites of digital monitoring, management, and automation (Macrorie et al., 2021). Yet the state of being or becoming smart can be forestalled, thrown into question, or subject to scrutiny (Leszczynski, 2016). Smart environments are meant to materialize through joined-up networks and systems to become coordinated, optimized, automated, and ultimately more “intelligent.” On one hand, “smart” can be a marketing argument or a concept to promote often-extractive technologies that promise environmental solutions even at a planetary level (McNeill, 2015; Söderström et al., 2014); on the other, it can refer to the joining up of technology, data, landscapes, social actors, and communities to address environmental problems (Foth et al., 2011). However, no specific technology, whether sensors or data dashboards, individually comprises smart environments. In this sense, the political effects and possibilities of smart environments do not settle into any single device or network, can vary across locations (Burns et al., 2021; Enlund et al., 2022), and can pose ongoing social problems (Rijshouwer et al., 2022). Rather than work within a fixed or universal definition of smartness, we consider how diverse speculations, materialities, technologies, and practices converge within and through increasingly digitalized environments in the making. This approach examines a plurality of smart forests as they are proposed, built, and transformed. Moreover, our intention is neither to advocate for smart environments nor to issue a condemnation of digital technologies. Instead, we seek to understand how digital environments are composed to make more or less livable worlds, and by what means.

By first reviewing “actually existing” smart cities literature along with “actually existing” references in social and political theory, we consider how to update and advance the “actually existing” analytic by revisiting the perceived rift between speculative and actual environments. While there is a broad range of smart cities research with which we could engage, we are especially interested in developing an in-depth investigation into smart forests drawing on the “actually existing” framework, as we suggest it can provide an approach for working through how new and emerging technologies materialize. In other words, it is the anticipatory and promissory aspects of technologies (Borup et al., 2006; Brown and Michael, 2003) that the “actually existing” analytic would address as they come into being and are iteratively developed. “Actually existing” can signal struggles to construct smart technologies and environments. Smart forests could be sites of increased community engagement and ownership. They could facilitate pluralistic modes of data collection and situated governance. Or they could facilitate extractive processes, which position forests as timber resources, carbon sinks, or ecosystem services (Gabrys, 2022).

Drawing on interviews with smart forest stakeholders, the second part of the article develops three examples of “actually existing” smart forests: the dis/connected forest, the present forest, and the proxy forest. These thematic areas demonstrate how inseparable the envisioning, making, and sustaining of smart environments can be, where smart forests manifest through pervasive or limited technologies, through in situ or remote research, and through different figures that stand in for forests. What these three themes further show is how the technologies that compose smart forests may be more or less evenly distributed, thereby conferring different benefits and (dis)advantages through collected and communicated data. Moreover, engagements with forests can become more distant through remote ecological research, remaking the practices and locations where forest encounters occur. Likewise, forests themselves are not self-evident and can be differently configured through the proxy entities that digital technologies seek to detect or document.

Our intended contribution is two-fold: First, we augment understandings of “actually existing” smart environments by working through smart forests as speculative, emergent, and built technologies and environments that converge with and diverge from smart cities. Despite their potential similarities, we suggest that forests and cities demonstrate differences in how smart forest environments come into view as speculations, practices, and projects (Nitoslawski et al., 2019; Prebble et al., 2021; Sarkar and Chapman, 2021). By focusing on the plurality of smart forests, we draw attention to specific power dynamics in designating, making and sustaining forest worlds (Devine and Baca, 2020; Marijnen and Verweijen, 2018; Peluso and Vandergeest, 2020). We especially focus on the intellectual trajectory of “actually existing” research to suggest that the “actually existing” does not signal one empirical condition but rather designates multiple and uneven social and technological arrangements. Second, based on our contribution from a smart forests perspective, we suggest that “actually existing” conditions do not exclude speculative visions, fantastical or otherwise. Instead, they take shape with and through ongoing eco-social concepts and practices. To this end, we suggest an approach for advancing “actually existing” analysis along this trajectory. Even more than empirical conditions counteracting fantastical visions, we propose the “actually existing” analytic can be updated and mobilized to examine how plurality, contestation, and democratic participation are at stake not just in the lived conditions of smart environments but also in the abstract and provisional contours of computational technologies as they shape and transform milieus. Concretely, this means it is as important to ask whose and which smart environment speculations are foregrounded, as it is to examine the material-spatial worlds that are built. We further propose that engagement with the plurality of speculations, material conditions, and practices is crucial for ensuring “actually existing” eco-technical worlds and relations are attuned to equity and environmental flourishing.

Actually existing as analytic

Key theorists of the “actually existing smart city” argue that the smart city often rests on a “nebulous idea” (Shelton et al., 2015: 13). As these authors suggest, the prevailing if hazy figure of technologically infused urbanism is characterized by digital hype, which obscures the practicalities of digital developments. The smart city can circulate within projects and plans that veer toward the fantastical and unreal. Such imagined cities are often friction-free computer simulations to be built on greenfield sites and operating free of cumbersome urban governance, unruly citizens, or resource constraints. The “actually existing smart city” makes an important intervention into these plans and visions, scaffolded as they are upon impossible actors and actions, to suggest a more “grounded” and empirical approach to studying smart cities as they are built and inhabited. The “actually existing smart city” provides a counteractive force to the “nebulous” ideas and plans of digital urban development.

By taking a more empirical and concrete approach, Shelton and co-authors critique the neoliberal and technocratic characteristics of smart cities as they are developed and built. Rather than the fantastical smart cities of Songdo, Masdar, or Living PlanIT Valley, these authors focus on Louisville and Philadelphia in the United States to analyze the more concrete and mundane realities of digital technology as it is integrated into urban life. Such an approach involves updating and addressing typically corporate smart city proposals by investigating cities in their lived realities. Based on their analysis of these more conventional smart city plans as they become built projects, Shelton and co-authors (2015) suggest that the “assemblage” of the smart city “bears little resemblance to the marketing rhetoric and planning documents of emblematic, greenfield smart cities” (p. 14). In contrast, their analysis serves as a critical corrective that has informed smart cities research over the past decade or so as researchers and urban practitioners engage with smart city proposals and projects with a more critical and nuanced eye to the lived worlds they would create, and not just the proposals they envision. Here, “actually existing” as a concept involves attending to the grounded-ness of smart cities, above and beyond “idealised but unrealised vision[s]” (Shelton et al., 2015: 14). This approach directs attention to the inequalities and disparities of urban life, which can be amplified or exacerbated in different urban developments (Shelton et al., 2015: 15), something that many plans gloss over or ignore. Hence, grounding smart cities is an important strategy for understanding “the promise and the peril of the smart city model” (Shelton et al., 2015: 22) by taking “a more nuanced and situated” approach (p. 14).

Such research seeks to ground, situate, and complicate the smart city. It also serves as the basis for further research into the “actually existing smart citizen” (Shelton and Lodato, 2019), which complicates different smart-city imaginaries of participation and civic exchange in the actual goings-on of urban democratic life. In turn, this approach has informed a broad range of smart cities researchers (Leszczynski, 2020; Rose, 2020; Sadowski and Bendor, 2019) who draw on Shelton et al. to examine the faltering, unequal, lived material conditions of digital urban environments. This range of smart cities research demonstrates how fantastical and often-exclusionary speculations about urban infrastructures, citizens, and processes can diminish possibilities for more democratic, just, and equitable urban-environmental exchanges, and how an over-emphasis on the speculative aspects of smart cities plans can neglect the lived aspects of smart urban worlds.

While the “actually existing smart city” has been in circulation for almost ten years at the time of this writing, the phrase and analytic of “actually existing” precedes smart cities research by several decades in different iterations. Shelton et al. (2015) craft their analysis by drawing on Brenner and Theodore (2002), who develop the notion of “actually existing neoliberalism” and consider the “contextual embeddedness” of neoliberalism, beyond the universal framework of neoliberal ideology. In this sense, the specificity of urban spaces becomes central to how neoliberal projects manifest. In addition, the concept and empirical strategy of the “actually existing” can be found on either side of this work in domains that exceed smart environments. Researchers have analyzed “actually existing” versions of urban regeneration (Ross and Vanolo, 2013), urban justice (Williams, 2017), platforms (van Doorn et al., 2021), and cosmopolitanism (Nash, 2009). Indeed, the list of “actually existing” scholarship is far too extensive to list in detail here, but also includes studies of identity, markets, globalization, the Internet, capitalism, socialism, austerity, jurisprudence, and much more. What is notable about many of these examples of “actually existing” conditions is that they often consider how abstractions and lived conditions are tethered together.

Moreover, many of these formulations are continuous with Fraser’s (1990) even earlier analysis of “actually existing democracy.” Fraser (1990) developed an analysis of publics and power dynamics, where “actually existing” involves attending to the abstractions that infuse publics and power, but which also exceed these abstractions through lived conditions. In developing this argument, she dissects Habermas’ (1989 [1962]) notion of the bourgeois public sphere by interrogating and reconstructing the concept of publics and considering whether it can “theoriz[e] the limits of actually existing democracy” (p. 57). As she articulates, this is an important critical theorization since by engaging with both abstract and lived publics it is possible to differently constitute publics in ways that are more equitable and just. This inquiry into “actually existing democracy” and the publics that would be imagined and experienced shows how concepts can enable or impede practices of differentiation and reckoning with power and inequality. Here, “actually existing democracy” becomes a site of theoretical examination and political possibility, where the conditions and actions of democracy require sharpened concepts to analyze and propose expanded opportunities for social-political life.

Rather than the “actually existing” serving as an invitation to ground speculations, with Fraser’s formulation, it poses a challenge to make ideas more robust and incisive in dialogue with lived conditions so that they have generative potential. Empirical examples are less the concrete correctives to unreachable abstractions. Instead, they show how concepts are co-constitutive along with lived conditions. As Fraser (1990) notes, the idea of the public can implement power over publics, influencing institutions and participatory practices that enable or undermine actually existing democracy (p. 57). What this suggests is not only that it is important to attend to “actually existing” conditions as they collide with something like rhetoric or discourse but also that it is critical to trace and transform the concepts that are percolating through social-political projects such as democracy, citizenship, and for the purposes of this study, smart environments.

While there is insufficient space within this article to engage with Fraser’s entire discussion of “actually existing democracy,” we draw attention to one additional crucial point in her argument. In analyzing the public as a conceptual distinction, Fraser demonstrates how abstractions never exist as singular or universal entities. Instead, publics are always multiple and competing, both in their abstract and lived instantiations (which we would suggest are also not easily divisible). This observation allows Fraser to critique Habermas’ idea of the liberal public, which does not account for other modalities of publics that are neither bourgeois nor masculinist. It also allows her to highlight the relations among multiple and competing versions of publics and counterpublics, which can in turn influence accepted standards of democratic life (Fraser, 1990: 61). The “actually existing” does not arrive at one singular version of publics, democracy, or cities, because there are many ideas, lived conditions, and political possibilities for how these could progress. Here, the “actually existing” is necessarily pluralistic because it signals the different modes of democracy that take hold in public imaginations, institutions, and forums, and which lead to ongoing contestations and power struggles (Fraser, 1990: 68). Plurality, in other words, is a crucial condition that the “actually existing” surfaces. Such plurality demonstrates how power, inequality, and contestation run through different and competing versions of publics. It also highlights possibilities for greater equity and participation by engaging with these pluralities. It is necessary to keep abstractions and lived conditions in play to continually work through and advance the conceptual distinctions that “the public” affords within “actually existing” conditions of democracy.

The “actually existing” smart forests that we investigate here provide cause for revisiting the empirical-speculative divide of smart cities discourse. We propose to move beyond this divide by considering how Fraser mobilized a plurality of ideas (about publics) and lived conditions as a baseline for “actually existing” scholarship. We further extend the “actually existing” analytic by drawing out the pluralism that is implicit in this conceptual framing. One idea about publics can never prevail because one way of experiencing publics is not tenable. Similarly, one speculative vision about smart forests will fall short because there are many ways that these environments are built and mobilized. In this sense, we are guided by Annemarie Mol’s (2002) formation of multiplicity in The Body Multiple, where she turns toward practices to engage with and study social worlds. Within the realm of social practices, the collision of speculative and lived conditions takes place in a multiplicity of ways. While in Mol’s case she attended to questions of medicine and disease, we suggest here that this turn toward practices and plurality that has by now become emblematic of science and technology studies research also allows us to revisit the “actually existing” analytic to consider how it can provide an analytic and technique for examining smart environments in process, as practical undertakings and accomplishments, as failed visions and power struggles, and as worlds and social practices in the making.

What does “smart” hold together across speculative and lived conditions, which together are meant to compose the “actual”? Within the broader context of digital developments, smart technologies perpetuate forms of engagement and development that would seemingly operate without material constraints, or that would circumvent social-political entanglements to realize a friction-free future. In this sense, we consider how “material” or “empirical” engagements with smart forests serve less as a corrective to abstractions and more as a proposal to work across co-constitutive speculative-material environments as they mobilize technologies, eco-social relations, and lived worlds. Smart urban visions are not free-floating but rather are connected to actors, projects, and power dynamics, which are all potential sites of intervention (cf. Goh, 2015). Our intention is not to find fault with “actually existing” smart cities research but instead to extend its conceptual purchase both by revisiting earlier “actually existing” research that maintains a focus on ideas as much as material conditions, and by developing this analytic in the context of smart forests. We suggest that this approach allows for a further development of how to undertake “actually existing” research into smart environments, while considering the pluralistic ways in which these take hold across ideas and lived conditions. This approach relies less on a narrative of falsification, or to establish that smart environment projects did not go according to plan. Instead, it seeks to consider the “actually existing” imbrication of speculative, lived, and uneven environments as they are developed under the sway of the digital. In the context of smart environments, this involves both attending to how certain technological speculations can dominate plans and proposals, while also analyzing the plurality of ways that technologies and environments are configured, built, funded, and maintained. The next section provides an analysis of three examples where “actually existing” smart forests surface as speculations, practices, and projects in the making.

Actually existing smart forests

As signaled in the introduction, smart environments can be in varying stages of development, from the built to the speculative. Smart forests might not even be designated as “smart,” since computational infrastructures in the form of remote sensing could be distant from forest sites (Bennett et al., 2022), or they could be installed as discrete testbeds or prototypes that are not scaled up to more extensive interconnected networks. The materiality of smart environments can vary significantly, with different digital infrastructures being more or less present for engagement by diverse actors. Remote geospatial data sets can travel to participatory apps for use by environmental defenders, or they can remain within remote centers where land management decisions are taken on behalf of forest inhabitations. The abstractions of “smart” further collide with the abstractions of forests, where questions about what a forest is can guide and inform policy, research, and land management (Chazdon et al., 2016; Gabrys et al., 2022). Computational technologies that circulate within smart proposals are often presented as more planetary-scale projects within big-tech plans. However, such technologies might also be in use for site-specific research. For instance, tree growth, soil moisture, carbon storage, heat abatement, air filtration, and biodiversity, become ecological conditions for monitoring, managing, and transformation with the aid of digital technologies. In other ways, Indigenous, traditional, and local communities use drones to inspect deforestation events, carbon offset schemes monitor land-use practices to create economic value from standing forests, sensor testbeds monitor drought and disease in stressed ecologies, and rangers respond to sensor-triggered alerts to inspect forest violations (Paneque-Gálvez et al., 2017). These use cases might even be enrolled into proposals for smart forests to make the technologies more accessible.

During a span of 18 months, we conducted over 50 interviews with stakeholders working at the juncture of forests and digital technologies. We used Atlas.ti, a qualitative data analysis software, to analyze our interview transcripts. We identified 129 thematic codes across interviews based on shared topics and word frequency. We then identified and narrowed down the code selection based on the “actually existing” concept. From this sifting process, we identified seven thematic codes: information asymmetry, contestations, presence, automation, datafication, surprises, and multiplicity. These codes signal how interviewees encountered and discussed smart forests across speculations, installations, development practices, unanticipated consequences, or unequal material conditions. A recurring theme within the codes was the disjuncture and uneven alignments across speculative forest technologies and the different smart environments they contribute to making.

While many shared topics resonated within these interviews, we then identified three instances where the “actually existing” smart forest materialized, including the dis/connected forest, the proxy forest, and the present forest, we discuss and compare material from six in-depth interviews with stakeholders in locations including Brazil, India, Iraqi Kurdistan, Malaysian Borneo, the United Kingdom, and the US. We highlight these six interviews to explore narratives about smart forests, which highlight the imbrication of speculative, lived, and uneven smart forest conditions. While there is not space here to explore all 50 interviews in depth, the six highlighted interviews draw attention to common themes shared across the broader data set. We work across interview locations to show how themes converge and diverge across diverse forest conditions and land practices and how they align with or challenge the “actually existing” indications of what a smart environment is or could be. These qualitative interviews resonated with our broader interview data, and presented particularly salient moments where smart forests were composed across speculations, forests in the making, and lived conditions, which become especially evident in interviewee narratives about smart forests. In our interviews with forest and technology stakeholders, we found there was neither a singular understanding of the contributions or limitations of digital devices, nor an agreed approach to smartness. In locations worldwide, stakeholders had diverse ideas about what might constitute a forest, as well as concerns about how to care for forest environments through the use of digital technologies.

The dis/connected forest

One of the primary reasons digital technologies proliferate in environments such as forests is to gather more data about environmental change. Within forests, deforestation, wildfire, disease, drought, and biodiversity loss are critical events that an increasing store of data—including real-time data—is meant to monitor and evidence (Lewis Hood and Gabrys, 2024). Yet within proposals for implementing more extensive monitoring networks or for synthesizing a greater number of data sets several contradictions arise. Visions about a near-total state of data capture through sensors can collide with concerns about inequality in accessing data (cf. Shelton et al., 2014). Sensor testbeds within scientific field sites can attempt to collect comprehensive data sets, while forest monitoring for community campaigns can focus on generating site-specific environmental data through sensors, drones, and remote sensing to mobilize legal cases (Millner, 2020). Those forests which are disconnected and without digital monitoring technology potentially will not register—or register in the same ways—due to a scarcity of instruments and data. Even within highly instrumented forests, “actually existing” forests are made evident through distinct variables and phenomena that technologies monitor. The smart forest materializes through data sets that are meant to establish patterns in changing forests. In this sense, the smart forest can be simultaneously speculative and “actual.” Forests are instrumented and monitored, but awaiting connections and consequences that are yet to be established within data sets for future use.

As part of our extensive review of smart forest projects, we met with a regional director of one major smart forests project to discuss the initial ideas, grant proposals, and eventual installation of digital technology in a forest at the Hubbard Brook Experimental Forest in central New Hampshire in the US. Hubbard Brook is a 3500 hectare field laboratory that was established by the US Department of Agriculture in 1955 (Campbell et al., 2021). This interviewee had extensive interdisciplinary training in ecology as well as philosophy and had a balanced yet optimistic view of what these devices could enable. Initial funding for this site’s smart forest initiatives came in the form of pieced-together funds, which allowed for the initial network setup. Within the funding applications, the forest organization proposed to set up a testbed that could demonstrate what was possible. As this interviewee notes:

“And it was really a ‘show me,’ it was a prototype, it was saying, ‘Gosh darn it, we can not only have data from one site, but what if . . . What if we had multiple wired forests talking to each other in real time.’ I mean it’s kind of mind-blowing, and we’re not there yet, but that was the idea.”

—Research Ecologist, New Hampshire

In our interview, we discussed the speculative potential of smart forest technologies, as well as the logistical challenges they present. While the interviewee describes the initial smart forest as a prototype, the development included sensor networks, data dashboards, and visualization and sonification tools. For this interviewee, the speculative contributions and insights that could be gained from wiring up forests with computational technologies align with many of the promises that smart environment proposals offer—comprehensive, real-time, and interconnected environmental data to measure and monitor critical ecosystems such as forests. The connected forest, in this sense, would be a more data-rich and manageable environment.

With this project and discussion, we found the speculative potential of the smart forest to be bound up with a more prolific generation of environmental data across the changing space-time conditions of forests. For this interviewee, the “actually existing” smart forest materializes as a highly connected and networked environment, from which new insights will surface through more sizable and interconnected data sets. Once digital technologies are installed and abundant data is generated, then advanced understandings of forest ecosystems are meant to arise. Speculation, in other words, emerges from connected and material conditions of smart environments, rather than as a vision preceding implementation. As this interviewee notes:

[. . .] we’re super excited about the plethora of other types of sensors and what they're going to tell us about our forest ecosystems.

—Research Ecologist, New Hampshire

The sensors and the data they generate become the source of insights that are yet to be fully realized. In turn, the formatting of data through dashboards, visualizations, and sonifications is seen to be a crucial step to making the most of these smart forest networks—in other words, technology begets more technology. While the actually existing smart forest was already interconnected and generating data, the potential insights and uses to be realized from these interconnected systems remained a site of speculation. In our conversation, we wondered how environmental sensor data was mobilized, and who might make use of it. In response to this question, the interviewee replied:

Who is using it? That’s to me the million-dollar question. These data are still pretty new. I don't know how to get people to use these data, I don't know how to bridge the divide between scientists producing Excel spreadsheets, and data graphs with millions of points, and the schoolteacher in town X, or the student, or the public manager, or any of those people.

Part of it is getting our data out there on data dashboards and really delving into this relatively new field of data visualizations.

—Research Ecologist, New Hampshire

This interviewee clearly expresses and understands that much more work is needed to make data relevant for expanded publics. The speculative potential of the smart forest is to be realized through further technology that presents “actually existing” data in different and potentially more accessible formats. The smart forest in the form of digital devices and networks has been built, but the data that it generates have yet to fulfill its speculative applications.

Here, it is impossible to extricate the speculative potential of the smart forest as a site of enhanced ecological insights from the actual forest replete with sensor testbeds and data dashboards that is in search of applications. In their operation, digital technologies continue to generate speculative uses that would require more technology to realize. The actually existing smart forest becomes a never-quite-finished project to realize through the ongoing installation of interconnected digital technology that generates abundant data for eventual analytics and insights. However, as social studies of science and technology have suggested, data do not necessarily become effective through interconnectedness and abundance alone. Instead, it is through the formation of facts, conditions of relevance, local practices, and power dynamics that claims can be made, evidence can be circulated, and responsive actions can materialize (Gabrys, 2016; Goldstein and Nost, 2022; Loukissas, 2016; Nadim, 2021; Walford, 2015). A more pluralistic approach to forests as eco-social relations and practices could be a way to rework the potential relevance of data beyond technoscientific expectations about the uses of data.

Not all forests are as highly instrumented as this one, however, and uneven data sets can generate considerable discrepancies across environments. In this way, another interviewee emphasized how connected forests can create information asymmetries in comparison to disconnected forests, thereby further excluding less connected communities. This interviewee based in a Brazilian environment, science, and technology organization suggested that connected forests could have impacts on disconnected forests. While this well-established scientific researcher was engaged in multiple remote-sensing and geospatial analysis projects in the Amazon, often working with big-tech companies such as Microsoft and Google, they worried that digital technology and data were not evenly distributed. As the interviewee notes:

Digital inclusion or exclusion components: that worries me a lot. Because you take the asymmetry of access to information, some communities are not included in this digital network and that’s very, I think, a problem because this asymmetry information is power so communities that have more information can succeed in a way [that] sometimes affect[s] other communities.

—Spatial Analyst, Brazilian Amazon

As they identified, while computational technologies might promise an abundance of data, this was an abundance that was only available to those communities, researchers, nations, or companies best situated to take advantage of these devices and networks. In the context of the Brazilian Amazon, as communities attempt to mobilize technologies for protecting their livelihoods and addressing issues such as deforestation and wildfires, this inequality could have immediate and palpable effects. The interviewee describes these inequalities and says:

I will give you an example and then you [. . .] will know perfectly well. There are some tribes in the Amazon that are isolated but other communities are not as isolated, they are using GPS, satellite data [. . .]. So you’re creating asymmetry of information so it’s important to take into consideration.

—Spatial Analyst, Brazilian Amazon

“Actually existing” smart forests are not just located directly within extensive digital networks but also in their adjacent and residual effects. Varying access to technologies and communication networks, uneven engagement with partner organizations, inequitable resource distribution, language barriers, and geographical isolation are just a few of the ways in which smart forest networks could collide with and produce information asymmetries across connected and disconnected forests (cf. Cifuentes, 2023; Wainwright and Bryan, 2009). The potential consequences for livelihoods are especially notable, as this interviewee further remarks:

This is an extreme case that really worries me. But I see places in rural areas that don’t have access, for example implementing technical systems to improve agriculture. There are a lot [. . .] of applications but [. . .] they are so isolated sometimes they cannot use this, they cannot benefit from this technology.

—Spatial Analyst, Brazilian Amazon

The power dynamics that arise from smart forest networks can create marked data inequalities depending upon whether communities are comparatively connected or disconnected, which this interviewee sees as a central concern for digital technologies in forest environments. They note:

And that’s why I heard from an economist that this concept of asymmetry of information, how it leads to power. I started to worry about this, how to make this more inclusive, I think that’s the crucial challenge for dealing with this type of technology.

—Spatial Analyst, Brazilian Amazon

In contrast to ideas of information abundance and insights that will flow from digital technologies and that resonate with typical smart city narratives, here every device and infrastructure brings with it the potential for inequalities within and among communities. Such speculative inequalities could have direct consequences for which communities can benefit from digitally informed land practices, and which cannot. If one community has access to deforestation alerts and can patrol its lands, but another does not, this could displace illegal deforestation practices to the community without digital monitoring equipment. Considerations of access, use, and mobilization of environmental data could, in turn, transform how smart forest networks are imagined and constructed. Here, the actually existing smart forest could generate inequalities for other forests that do not have access to digital resources. While not officially “smart” such disconnected forests could experience digitally aligned impacts that do not readily feature as the development of a smart environment, but rather as the fallout from adjacent smart forests. The “actually existing smart forest” can therefore create speculative and actual harms for more disconnected forests.

A forest might only be episodically “smart,” moreover, where discrete and time-limited uses of digital technology could enable data collection. For many Indigenous, traditional, and local forest communities, environmental data collection takes place as part of projects for informing land practices, making claims and protecting land, and raising legal cases when infractions occur. In this sense, specific data can be mobilized toward particular projects and objectives, rather than delivering speculative insights through an ongoing abundance of data. Another interviewee working on conservation technology in India described this as a central feature of their organization. In their work, drones, remote-sensing data sets, sensors, and data analytics came together often on a case-by-case basis to support campaigns, legal challenges, risk assessments, and policy. As this interviewee describes:

Going back and turning to this question about partners. We, you know, again, we started and worked with people like Deloitte and PwC as well on risk assessment in this particular issue, right. And to give an example of that work, we convinced them that actually looking at biodiversity risk from the perspective of how biodiversity affects the investment risk of their projects. You know, assessing that and using that to make better decisions which would both make sense from a business perspective as well as for conservation [. . .].

So large-scale assessments of road networks across the country, looking at where their clients were investing in road networks [. . .]. It was a risk from the fact that biodiversity exists over there.

—Conservation Technologist, India

Here, the connected forest is a momentary installation for gathering data to address environmental problems and building community networks, especially for addressing road-construction projects. With this proposal, the interviewee suggests that taking a broader view of biodiversity data, as well as making connections with understandings and projections of risk, could allow for a different analysis of the impact of road networks on the land. When deploying digital technologies to monitor new road projects, different connections can be made across data sets by considering how environmental problems are interlinked. In this sense, data is less about abundance surfacing scientific insights through comprehensive digital infrastructures, and more about a project of mobilizing data to make a case for joined-up environmental practices and policies. While digital technologies might be in constant operation or episodically deployed, they gather their effectiveness through targeted uses. The “actual” technology is underwritten by its speculative potential, where data have the potential to evidence environmental harms. “Actual” technologies are less exclusively the grounded example of how smartness takes hold, and more expansively the multiple eco-social and eco-technical practices that could be activated.

As one iteration of the “actually existing” smart forest, the dis/connected forest is not characterized by pervasive or limited digital monitoring, but instead by these shifting constellations of how technology, milieus, speculations, and interventions are co-constituted and mobilized. The “actually existing” smart forest is not a single environment, empirical reality, or lived condition, but instead a pluralistic range of practices, installations, speculations, inequalities, data-in-waiting, and site-specific projects to document forest changes and mobilize data and communities. These different smart forests can be held in play, rather than reduced to one universal framework, to consider the contributions that can be made by each, as well as the concerns that different approaches might raise or encounter, as the next section further describes.

The present forest

“Actually existing” would seem to suggest a smart environment or forest that is most present, self-evident, or material. However, digital technologies co-constitute forests in distinct ways. With many digital technologies, forests become present as stores of carbon and biodiversity, and as sites of environmental crime and dispossession. Sensors, satellite imagery, and camera traps might respond to some environmental concerns and make forests present in response to these agendas, but they can also exclude other knowledges and values to which sensors might not be responsive (Gabrys et al., 2022). In this way, a smart forest could most frequently be envisioned and developed as one that responds to the objectives of environmental scientists and policymakers, who seek to preserve forests as carbon stores or biodiversity reserves (Gupta et al., 2012; Nel, 2017). In making the forests present, the “actually existing” analytic directs our attention to the multiple transformations that take place in and through technologies.

For ecologists, digital technologies can present a challenge for their relative connection or distance from forests. Technologies can remove researchers from the field, but they can also remove the field from the technology. For example, estimating the number of birds as a critical indicator for ecosystem biodiversity is difficult in high biodiversity environments like tropical forests (Metcalf et al., 2022). Traditionally, “point counts” are used as a rapid assessment for capturing bird biodiversity in a particular area (Leach et al., 2016; Metcalf et al., 2022). This method is done by recording all the birds that are seen or heard at a specific location and time (Klingbeil and Willig, 2015). Nonetheless, this approach poses challenges in regions with rich biodiversity, where data overload can overwhelm observers and lead to the omission of certain species. Discrepancies in observers' expertise can hinder accurate survey comparisons, and the mere presence of observers in the field may alter bird behavior. Furthermore, the costs of extensive in situ fieldwork can also be prohibitively expensive (Leach et al., 2016).

In response to these problems, Autonomous Recording Units (ARUs) have been developed as a supplement due to their ability to reduce several types of bias associated with the point count method and to enable consistent data collection among surveys that facilitate the improvement of species detection and richness (Klingbeil and Willig, 2015). ARUs collect and record data about the habitat, behavior of the birds, and changes in the sound without the presence of a human observer (Symes et al., 2022). While these digital technologies can address some of the point-count problems, Klingbeil and Willig (2015) show that the accuracy and richness of biodiversity estimates from the point-count method was greater than the from ARUs. For instance, the technology does not seem to detect organisms such as woodpeckers. One interviewee who had set up an acoustic monitoring project in Malaysian Borneo spoke to these challenges across automated, remote, and in situ technologies and practices, and noted:

But obviously there’s subtleties to the data that someone doing the point count gets that an automated approach wouldn’t [. . .] such as they might know the sex of a call, for example, or the breeding stage of a call, and you might not have implemented that in your algorithms. You know you talk to a field assistant who’s been doing a job like this for six months [. . .] and they say, “Well actually, how about this?” This, that, and the other that the scientist who’s coordinating the project maybe hasn’t come up with, and it’s that intuition that you gain from just being in the field.

—Bioacoustic Sensing Researcher, Borneo

Here, experience gained from being in the field informs how the digital technology is mobilized. Rather than insights flowing from the technology, these must be gained through a complex negotiation across prior field experience, data, algorithms, and ecological methods. The “actually existing” smart forest can be differently composed through the presence or absence of technologies, researchers, and variables sensed or detected. This is not a singular smart forest, but one differently activated through eco-social practices that could value some species or ecosystems over others. Different iterations of forests materialize across in situ, remote, instrumented, and/or automated encounters. Describing how intuition reshaped the research, this interviewee emphasized:

It’s so invaluable just to go and see the field site for a day or two as much as you can, because you just get some intuition for the kind of place that you’re surveying and the data you’re working with. So, if you go to automated you can lose that intuition and ignore quite obvious things that would be apparent if you go there.

—Bioacoustic Sensing Researcher, Borneo

This interviewee suggests that automated digital technologies could erase or supplant insights that can only be gained through field experience, described here as intuition. With automated observation technologies continually taking measurements, there is the potential that these devices could even create a new baseline for common sense. Within this interviewee’s analysis, there are at least two types of observers: one who is present in the forest, having a knowledge of local conditions; and the second who is developing algorithms and interpreting data but is remote from the field. Both observers are connected to a different yet similar forest (cf. Awan, 2016). Whether these different observations and practices intersect is not always guaranteed. “Actually existing” forests are to be monitored for their biodiversity. However, different conceptual formations of biodiversity, as well as evaluation technologies and practices, can make different types of biodiverse forests present (Westerlaken, 2024). A multiplicity of forests materializes here, which could complement each other or serve as the basis for contestations in spaces of environmental management and governance.

These different registers of forest presence are also evident in the Hubbard Brook forest site discussed above. This field laboratory was developed to detail the relationships between streamflow, vegetation, and the chemistry of the precipitation (Campbell et al., 2021). Hubbard Brook has one of the longest field records of soil frost, where their technicians have measured tactile frost depth manually since 1956 (Campbell et al., 2010). Manual measurements of soil frost were made at a single point at each location at weekly time intervals (Campbell et al., 2010). These measurements led to variations in the observed soil forest depth and disagreement between the predicted and observed values due to errors during the manual measurements (Campbell et al., 2010). Responding to the adoption of the limitation of manual measurement, the Hubbard Brook interviewee highlighted the contrast between manual measurements and the use of digital sensors:

Instead of having weekly measurements, I can have measurements every hour, every 15 minutes, every minute. And I don't have to depend on a technician going out to do these. And then if we wire it up effectively, we can have these transmitted to our base stations through telemetry, and we can have these measurements at the palm of our hand.

—Research Ecologist, New Hampshire

Forest ecologies become present here through a greater frequency and granularity of data, as well as by transmitting distant events to researchers in their base stations. Such an approach promises greater immediacy and accuracy. Emphasizing how transformational this digitally informed approach to environmental science can be, our interviewee continued:

So, realize that the old way of doing business was collecting measurements, putting it in spreadsheets, somehow uploading those into databases, and if we were lucky, we might get the data six months, a year, 18 months later. Now all of a sudden, we had the promise of having real-time data, big data like the urban data [. . .].

—Research Ecologist, New Hampshire

Contrasting the “old way” of measuring forests with digital approaches, this interviewee draws attention not just to the contrast between in situ or remote experiences of forest events but also to the multiple transformations that take place to data depending upon the formats and frequencies in which it is gathered and processed. The episodic and partial qualities of hand-gathered field data can make for a less expansive data set than a digital sensor network. In this interviewee’s perspective, wired forests can even approximate cities by generating big data sets with streamlined and automated processes. “Actually existing” smart forests would then be differently composed and materialized, depending upon whether data sets are more immediate and expansive, the variables used, and the absence or presence of researchers, among many other factors.

While data collection and analysis might be common practices, their implementation can take place through different tools or techniques that differently constitute forests. This can, in turn, also contribute to different understandings of forest events based on proximity, experience, interpretive techniques, and further responses that these could generate. Are remote scientists better placed to describe biodiversity in tropical rainforests, or are forest dwellers who live alongside these organisms and ecosystems? Whether wired or not, the multiple observations and data sets that forests generate can make forests present in different ways, both as speculations and lived sites. They are “actually existing” as distinct sites due to measurements and data protocols, the scientific objectives to which these respond, and the analogue or digital practices that evidence and solidify these versions of forests. In this sense, different eco-social and eco-technical practices can make a multiplicity of forests present. This in turn can also inform the power dynamics of who can or cannot speak for or about forests.

The proxy forest

If the previous two examples show how digital technologies are often deployed to connect up forests and co-constitute them through registers of data dis/connection and presence, this section considers how the very indicators of what counts as a forest can be subject to scrutiny, depending upon which entities stand in for forest health or ecosystem integrity. While the concepts of “smart” and “forest” help to investigate a range of power dynamics between speculative and lived conditions of digital and forest environments, our interviews also revealed how the “actually existing” smart forest is co-constituted along with multiple other entities that would ordinarily not count as a “forest.” Within numerous interviews, we identified multiple proxies that served as approximations of forests. These proxies reworked the focus on forests and technology, and showed how these are not fixed or singular concepts or relations. Forest worlds, in other words, consist of other entities that can reconfigure “actually existing” forests as they are usually designated.

One example of how digital technology indirectly produces such an expanded forest world emerged through the medium of water in the context of the Brazilian Amazon. Here, the Brazilian senior researcher whose insights are included above explained how issues of deforestation and carbon emissions can be challenging to communicate to general publics. Water, on the other hand, is mobilized as an idea that can more directly reveal issues that help people to understand the risks of climate change:

I think that the challenge is how to communicate [environmental issues] to ordinary people [. . .]. That’s really the challenge. And [we] start[ed] to work with water because water is connected to [. . .] climate change.

People don’t understand much why we are worried about the Amazon becoming our source of emissions, instead of a sink. This concept, it’s hard to be understood by ordinary people. But water: it’s straight. It strikes people directly, and there you can connect water forms to other environmental issues. That’s the strategy at least that we are trying to implement now.

—Spatial Analyst, Brazilian Amazon

As this interviewee suggests, by focusing on water rather than the usual dynamics of forests and carbon, it could be possible to raise public awareness about the risks of deforestation and climate change. In this interviewee’s assessment, it can be challenging for publics to understand or engage with forest data and technologies. Indeed, forests are almost too abstract in their “actually existing” form since carbon-based ways of documenting environmental change often do not adequately connect to public concerns. Instead, when traditional, local, and Indigenous people are faced with continued degradation of the Amazon through diminishing water supplies, the risks of climate change become more immediate in the context of ongoing drought and water restrictions. This example shows the “actually existing” analytic is more than a self-evident description of forests as trees, but also includes an extended range of forest proxies that could activate different eco-social relations and practices. Water is more immediate as a lived condition and experience of scarcity, which simultaneously activates an anticipation of forests under threat.

In the context of the Brazilian Amazon, digital technologies and data infrastructures have been sites of global and local political tensions for decades (Sanches et al., 2021; Vurdubakis and Rajão, 2020). While these extensive remote sensing data sets, sensors, and data dashboards enable this researcher to understand environmental change in forests, such digital technologies for carbon offsetting and deforestation are becoming increasingly complex. Technological developments for deforestation, carbon offsetting, or tree planting are now being co-opted for corporate growth or political gain (Urzedo et al., 2023; van der Hoff et al., 2018). This interviewee worked on an online platform with other partners that uses satellite data to create a timeline that visualizes how above-ground water, specifically in lakes and rivers, has disappeared since the 1980s. Here, “actually existing” smart forests show up in the medium of water, as a more immediate connection and point of both speculated and lived concern. Digital forest technologies enabled the creation of evidence to show diminished water supplies, as well as the consequences for Amazonian forests. Such technologies are conceptually attuned to the idea that water will be more aligned with public concern. In other words, by using water as an entity that affects both people and forests, in both its lived and potential understanding, the interviewee sought to shift focus and raise public awareness about the need to restore and protect Amazonian forests. This example shows how the development of smart forest technologies can be grounded in connections across people, water, and forests, as well as a more vivid if speculative sense of increasing water shortages.

While water can stand in for forests in the Amazon, in the Midlands of the United Kingdom another interview participant proposed that the very notion of a forest needed to be rethought in the context of the UK landscape. In these environments, they argued, carbon capture and biodiversity restoration cannot take place exclusively in forests as usually understood in the sense of the Amazon (although the cerrado is an important if overlooked aspect of these landscapes) but must instead include mixed woodland and meadowland. This participant describes the importance of the concept of a wood-meadow:

It depends on your definition of forest, to be honest, if you look at the old English term forest, it was almost a parkland landscape that was used for hunting, so it wasn’t necessarily full of trees in the UK, so you’ll find that there are forests in the UK that are not just trees. [. . .] One of the terms that’s coming to the fore is the use of the word called wood-meadow, so that is, if you like, the concept of a mixed woodland and meadowland, and the aim there is to capture carbon but also address the biodiversity loss that we’ve had.

—Biodiversity Social Enterprise Founder, UK Midlands

The notion of a wood-meadow shifts the attention from forests toward more mixed landscapes that involve a variety of landscape restoration practices. This wider speculation about what a forest involves can subsequently inspire different ways of calibrating technologies to monitor such sites. For this participant, automation processes enabled by digital technologies—particularly satellites—play an important role in determining where restoration can take place across these landscapes:

There’s about 40 layers of GIS information that are relevant to a plan such as this [. . .]. So, one of the things technology could do is automate that process, so we put a shape around a new county, push a button, and a day later, depending on how much data is being crunched, a map would be produced, and I think that would be entirely feasible, in fact, I know it is. It would need a check, and it would need verification, and it would need consultation, and it would need the stakeholders to adopt it, but it would be a good start and it would save everybody quite a lot of time and money.

—Biodiversity Social Enterprise Founder, UK Midlands

The interviewee proposes to apply automation technology not only in the context of tree-dominated landscapes, but across a variety of other restoration landscapes, from peatlands to grasslands. They highlight the importance of restoring soils and wildlife to create a more balanced, native, and mosaic landscape. Throughout the interview as well as in the quote above, it also becomes clear that such an approach involves several stakeholders and networks that influence restoration projects, leading to new power dynamics that may largely become directed by automation technology. Here, restoration informed by automation remakes the landscape matrix into a woodland meadow. The “actually existing” smart forest does not settle into a tree-dominated landscape, but instead is reconstituted as a patchwork of different sites that can be variously enrolled into restoration objectives.

As these projects show, forest environments can be differently constituted through digital technologies that intersect with a broad set of ideas and speculations about what forests are or should be. Different entities become vehicles for mobilizing attention toward forest protection and restoration. In Iraqi Kurdistan a different entity has also materialized to mobilize people in conflicted forest and mountain landscapes. Here, camera traps were used to identify wildlife in an area where many forests are lost because of fires caused by drought, but also because of human-made fires and attacks in conflicts between Iraq, Iran, and Türkiye. While monitoring forest events in 2011, one of the camera traps documented a Persian Leopard, an animal that was considered extinct in this region. The Persian Leopard became an important entity for sparking negotiations about land protection and freedom for local communities. As the interviewee based in this region noted:

[. . .] everybody was very excited [when we recorded the leopard], because then you could use this charismatic species to bring more projects. And using it as like an umbrella species to receive more attention, not just locally, but also internationally.

[. . .] Iraq’s forests are lost by at least 50 percent—more than 50 percent—compared to 20, 25 years ago. And these are related to natural forest fires or human forest fires related to war and raids along the borders, but especially recently with continuous bombardment and raiding on both the Turkish and Iranian borders. And mind you, that these border areas are really important for wildlife, because they allow the migration of these animals, especially an animal like the Persian leopard.

—Wildlife Conservationist, Iraqi Kurdistan

With camera traps, conservationists were able to visually document the leopard. This evidence in turn could raise international attention not only for this particular species but also draw attention to this conflict and the struggles of local Indigenous populations in their territories (Horeni et al., 2022). Since then, camera traps have been used in different regions to document the leopards. The Persian leopard’s range spans across forests and landscapes in 11 countries, and camera trapping has become more popular in these regions. Digital documentation of the leopard has also stirred up cross-border politics in Türkiye, Afghanistan, Russia, Georgia, Azerbaijan, and Armenia (Goodall and Tehrany, 2022). The “actually existing” smart forest is mobilized through camera traps, image data, infrequent leopard sightings, lived experience of forest loss, and the speculative promise of greater forest protections in a landscape under threat.

However, the use of forest technologies to document leopards can have other social-political impacts beyond forests. In 2018, the Iranian authorities attempted to charge eight staff members of the Persian Wildlife Heritage Foundation with “corruption on Earth” (which carries a death penalty) for the charge of spying against Iran with camera traps using foreign-made technologies and exchanging camera footage abroad. Kavous Seyed-Emami, academic and director of the organization, was arrested for the use of camera traps and died in jail (Goodall and Tehrany, 2022). Through international intervention, the sentence of the remaining seven prisoners was reduced, but the conservationists served 6 to 10 years in prison before eventual release. Camera traps are a key technology within smart forests but the images they collect and the locations in which they are placed can lead to present or future impacts within and far beyond the forests in which they operate (Adams, 2019; Benson, 2010; Sandbrook et al., 2021). As a more-than-human entity moving across different forests and borders, and showing up in unexpected places, the leopard shapes territorial conflicts and political power dynamics. The forests through which these leopards roam are transformed by documentation that would alternatively protect their habitats, prevent knowledge sharing, or enable further forest protections.

The proxy forest could materialize through digital technologies, but it also can upend usual ways of encountering the “actually existing” smart forest. Besides the numerous ways in which smartness is conceptualized in forests, including as connected sensors, digital platforms, networked satellites, algorithmic operations, or data infrastructures, the notion of the forest moves beyond its typical understanding as digital technologies interacting with a tree-dominated landscape. Devices such as remote sensing technologies and camera traps monitor proxy conditions and organisms, which stand in for and mobilize forest environments. Proxies are both within and transformative of forest conditions, signaling speculative and lived forest conditions. “Actually existing” smart forests become evident and relevant within vastly different conditions, political contexts, and contestations over forest speculations and lived conditions. The “actually existing” smart forest becomes present as a much different conceptual distinction in each of these sites, leading to a more pluralistic understanding of how digital technologies materialize in forest environments. Here, the reconstitution of speculative and lived forest conditions also suggests different ways to protect, restore, expand, and regenerate forest worlds.

Discussion and conclusion

Within “actually existing” smart cities literature, the suggestion that speculations are distinct from built environments prevails. Nevertheless, this article revisits this split and suggests that it is important to attend to the co-constitution of diverse visions and lived worlds when analyzing smart environments. We engage with the important and influential work of “actually existing smart cities” to better understand how plans and proposals for smart environments are built in uneven and faltering ways (Houston et al., 2019). Yet we also revisit the longer intellectual development of the “actually existing” analytic to attend to the plurality of smart environments. This expanded “actually existing” analytic allows us to grapple with the speculative and built conditions of smart forests, not as binary conditions or correctives to one another, but instead as interwoven processes and practices that make and remake forest environments. This analysis refocuses attention on the speculations, abstractions, and cultural narratives that influence smart environments (cf. Esposito et al., 2021). It also shows how power and politics pervade smart environments often to entrench practices of imagining, making, and creating eco-technical worlds. From smart cities to smart forests, these imbrications of speculative, lived, and uneven environments inform “actually existing” conditions.

Within the broader context of numerous environments becoming “smart,” this article has specifically investigated “actually existing” forests to consider how digital technologies influence and remake these environments. In our investigations into smart forest technologies, we have found that the “actually existing” smart forest circulates and takes shape across speculations, plans, practices, and in situ technology installations. As the six interviews in different locations worldwide suggest, forests and technologies can align with narratives similar to smart cities. Smart forests can be sites of social-political struggle over how technologies and environments are envisioned and implemented; and they can signal inequitable and problematic big-tech development projects that disenfranchise or dispossess inhabitants.

In another register, smart forests create different configurations of environments, technologies, and eco-social practices. While sensor networks, satellites, lidar, GIS, data dashboards, camera traps, and automated recognition units circulate in these projects and even at times form networks and infrastructures that could be understood as “smart,” the social-political and material conditions of these different environments unsettle any singular understanding of these dynamics. As we have demonstrated, multiple “actually existing” smart forests exist. However, some of these become more prevalent and realizable because of the power dynamics and circulation of abstractions that would value some forest worlds over others. Different technologies could enable community-engaged data apps and practices, while others could create conditions for further extraction. Digital technologies are activated within and in relation to forest conditions. If the primary visions of forests and digital technologies are directed toward carbon accounting and timber optimization, for instance, this will inevitably impact possibilities for other eco-technical worlds and cosmologies (Gabrys et al., 2022). The plurality of these forest worlds can also create different configurations of environments, technologies, and eco-social practices.

In analyzing “actually existing” smart environments more generally, we are not advocating for the recuperation of “smart” but rather proposing a technique for pluralizing approaches to these eco-technical worlds. By attending to the plurality of these forests and forest speculations, we seek to surface the power dynamics that support some eco-technical worlds, often to the detriment of others. We also work toward reconfiguring relationships across environments, technology, and social life toward greater equity and multiplicity. This re-engagement with the “actually existing” analytic highlights the speculations and pluralities of proposed, actual, or reconstituted smart environments to reconnect to the importance of abstractions while also showing how environments and technologies propagate in multiple forms that influence lived experiences. Such an approach engages with the diversity of concepts and worlds through which smart environments take hold and become sites of governance and regulation to identify and implement possibilities for greater environmental flourishing. For this reason, we propose to expand and extend the “actually existing” analytic by engaging more fully with these trajectories across speculations, action, power, collectives, and worlds, where “actually existing” eco-technical worlds are always multiple.