Digital twins of the ocean: Wet environing media and marine futures

Abstract

The EU’s Digital Twin of the Ocean (DTO) is presented as a coherent, high-resolution, multi-dimensional, multi-variable and near real-time representation of the ocean that integrates new data sources with advanced modeling, artificial intelligence and high-performance computing. The two-way exchange of information between marine ecosystems and the digital twin is intended to create a feedback loop between the digital and physical realms that is emblematic of environing media. The EU’s vision is that the DTO will empower citizens, inform politicians, support a blue economy, and improve protection of the marine environment. The sociotechnical imaginary of the DTO presents a narrative that a balance between sustainable exploitation and conservation can be had through the use of sophisticated digital technology. But what is at stake in this technocratic control over the world ocean, and where does it come from? Whose interest will digital twins ultimately serve? Understood in its historical and environmental context, the DTO caters to the aims of sustainable development, but risks veiling continued unsustainable development and growth under the guise of new digital technologies. We use the theoretical lens of environing media and sociotechnical imaginaries to critically unpack this narrative and its historical contingencies, and show how difficult goal conflicts are systematically glossed over through a veil of datafication and technological development.

Introduction

Ocean protection as well as ocean exploitation are intricately connected to the subjection of the ocean to an accelerated pace of mediation and datafication since the mid-twentieth century, and especially in the past two decades. New technologies have made many forms of ocean knowledge increasingly reliant on computation and data processing in digital infrastructures. The rapid rise and deployment of digital technologies for ocean research and management, including machine learning or artificial intelligence (AI) and widely distributed autonomous sensor networks, is informing ideas for integrating these data in so-called “digital ocean twins.” Such a digitalized ocean is both the object and outcome of human decision-making; digital technologies are profoundly reshaping our perception of marine environments and the possible ways to approach, use and manage them. This brings the issue of governance to the fore, as the reach and potential control over the marine environment is seen as dramatically increasing with access to new types of data flows and raises fundamental questions about the relationship between these new data and democratic influence, equity and justice in the allocation of resources for human needs and their environmental impact. Will digital twins of the ocean propel sustainability at sea as its developer claims, or is it rather a technocratic control fantasy that will feed into and superpower the blue acceleration (Jouffray et al., 2020)?

In this paper, we offer an in-depth analysis of the discourse around digital twins of the ocean understood both as a case of a sociotechnical imaginary that is defining ocean futures, and an instance of “wet environing media” that is effectively changing the human-ocean relationship through increased technical mediation. The paper aims to contribute to a critical understanding of the role of digital media technologies in recognizing and addressing the inherent conflicts between economic development and environmental protection with regards to the global ocean by approaching digital ocean twins as collectively held, institutionally stabilized and publicly performed visions of desirable futures attainable through advances in science and technology (Jasanoff, 2015). We also aim to place the digital ocean twins under current development into historical context, to enable a view of them not as essentially novel inventions but as a continuation and iteration of a long history of developments and visions for knowing and using the ocean for human purposes. Our main focus is the European union’s digital twin of the ocean, the EU DTO, which is the most advanced digital ocean twin-initiative so far and at the time of writing is being presented for the first time. The fact that the EU DTO has not yet been implemented means that we are restricted to analyzing the surrounding discourse and drawing out the policy implications rather than being able to study a technology that has been in operation over time. We historicize the EU’s presentation of this technology by viewing it as an act of environing and show how it risks reinforcing power relations and increasing extraction while glossing over environmental impacts through a veil of sophisticated technology.

Our analysis draws on recent theoretical advances in environmental media theory to analyze interactions between environing media and marine environmental sciences and policy. The role of digitalization in achieving ocean sustainability is often highlighted in policy and public discussions but has not been extensively or critically studied so far (Bakker, 2022). A better understanding of how the ocean digitalization process is unfolding and what new power relations are created, particularly with regard to public-private relations and the tension or even conflict between extraction and conservation, is needed. Our aim is to move beyond a naïve and uncritical understanding of goals and measurements as transparent and inherently good, toward a critical and integrated view of how data is gathered, processed and transmitted in chains of mediation that produce epistemic truths and views of the marine environment.

In the past decade, a new field of study referred to as blue humanities has aimed to shift attention from a terrestrial bias in many areas of scholarship to an aquatic which does not shy away from the fact that the ocean covers more than 70% of the Earth’s surface. The introduction to a special issue on “hydrocriticism” proposed to conceive of this increased interest in the seas as an “oceanic turn” (Winkiel, 2019) while another introduction that focuses on blue humanities in relation to science studies urges us to consider how “questions of scale, temporality, materiality, and mediation emerge in aquatic zones and modes” (Alaimo, 2019). This paper contributes to this field by zeroing in on the most recent development in ocean knowledge and governance in the construction of digital twins, understood through the theoretical concept of wet environing media.

The paper begins by discussing the development of wet environing media, which provides the theoretical framework for the analysis. We then introduce our case study, detailing the emergence of digital ocean twins as a novel approach to ocean science and governance, and explain in particular the stated aims and visions associated with the EU’s DTO. In section four, we historicize this new development by placing it in relation to a long history of ever-increasing collections of marine data that in some ways the DTO can be seen as a culmination of. In relation to this history, we discuss the promises made by the DTO with regard to how the ocean can be known and discuss limitations of data not acknowledged in the DTO discourse. In section five, we shift focus from how the DTO purports to know the ocean to how it envisions itself as situated in between efforts to protect and exploit the marine realm, or, differently phrased, its role in the sustainable development of the ocean, where we suggest that the DTO functions as a technological veil that hides persistent conflicting views on the human-ocean relationship, and contributes to obstructing, rather than informing, informed and accountable debates of contested aims. Finally, to conclude we point to some different ways in which digital ocean twins, if reformed and reformulated, could play a part in mediating meaningfully between marine sciences and policymaking.

Wet environing media

Environing media is a theoretical concept designed to uncover the technological preconditions for global anthropogenic environmental change. What we know about the environment—its epistemologies—are deeply related to the media technologies at hand, which form the conditions of possibility for environmental discourse (Wickberg and Gärdebo, 2022). At the heart of this process, which reaches back to the early modern era but took off at a radically intensified scale in the postwar period, is a complex feedback loop between knowledge and human impacts on the global environment. This dynamic can be exemplified through the growing technosphere—Earth’s sixth sphere consisting of all human technologies combined and function like the lithosphere, hydrosphere, cryosphere, biosphere or atmosphere (Haff, 2014), which essentially means that the sum of all technology can be understood as a self-regulating system. Following the unprecedented technological developments of computation associated with the second world war and its need for nuclear bombs, scientists as well as others have increasingly been changing scales from direct observation to interaction at a mediated level, to the extent that we can now speak of a mediated planet (Wickberg et al., 2024).

The theory of environing media targets the epistemological construction of environments, and shows how chains of mediation can be analyzed from the data gathering in distributed sensor networks, through the modeling based on this data, which give rise to epistemic concepts such as climate change or global mean surface temperature, and then get translated to the climate policy, actual politics and the general public understanding. In the case of the ocean, similar chains occur when measuring sea-level rise, ocean heat content or ocean acidification. It is through aggregating and scaling up that a phenomenon becomes knowable, but in this process the local and contextualized inevitably gets lost, which has profound political implications in that it often serves dominating technocratic interests while disregarding the reality of local communities, be they human or non-human. Using the verb form, environing signals a process-oriented understanding of the environment as something less stable and subject to changing understandings over time. The concept also responds to a mutual need for more robust and integrative theorization in the respective fields of environmental humanities and media studies of these matters, the combining of which we signal here through the notion of “wet environing media.” As pointed out by Miles (2023) with regards to precision agriculture, farming has become successively subject to extensive mediation over the past decades (cf. Wickberg, 2023), and this media aspect of human land use is largely overlooked as such. Something similar holds true for the changing use of oceans, in which datafication and mediation has brought ever deeper realms into our understanding but also fueled a blue acceleration (Jouffray et al., 2020).

The concept of environing was developed in environmental history over the past decade to reconsider how environments are made, against a general understanding of a stable and unchanging nature, disconnected from human history. The word appears already in the early modern era and refers to the encircling of a piece of specific land for the purpose of farming, but then faded from use. Sörlin and Warde (2009) proposed reintroducing it through a “history of environing, the study of how all human activity demarcates and generates an environment, an outside that haunts the space in which people choose to act,” informing studies of different environing processes over different historical periods. Sörlin (2013) defines environing as “what people do when they transform nature into environment,” and notes further that all of these ways of changing the environment that humans intentionally and unintentionally pursue create new historical relationships between land, ocean and people that are both social and environmental (Sörlin and Wormbs, 2018).

The essential means for environing can productively be understood in terms of media, which also connects this framework for environmental history to an emerging broader notion of media, understanding the concept to include elemental and planetary processes, as well as smaller scale change. On one end of the spectrum of such media are apparent physical technologies for changing and observing the environment, such as contemporary ocean sensing networks, while at the other end of the spectrum of environing are human imaginaries and desires which, often attached to the potential of media technologies, come to shape perceptions of desirable futures, and then shape technological development toward certain environmental futures in a constant feedback loop understood here in terms of the notion of sociotechnical imaginary, which Sheila Jasanoff (2015) defines as “collectively held, institutionally stabilized, and publicly performed visions of desirable futures, animated by shared understandings of forms of social life and social order attainable through, and supportive of, advances in science and technology.”

Following Steve Mentz (2020) suggestion that we should understand modernity in terms of a “wet globalization” to reflect the fact that the fundamental technology enabling many of the essential historical changes was oceangoing ships and cargo, we propose that digital twins of the ocean be understood as “wet environing media” that are in the process of changing our relationship to the world ocean by expanding the human reach for exploitation, extraction and pollution into the marine realm. This conceptualization also recognizes the need to displace and challenge terrestrial bias in media studies and attend carefully to how the ocean environment produces different forms of mediation (cf. Jue, 2020). Understood as wet environing media, digital twins of the ocean are quickly coming to support visions of expanding the reach of human endeavors to depths hitherto unknown, but it does so in a specific systems-oriented and physics-based approach, which ignore the complexity of biological processes and ecosystems and feeds back into a resource hungry conception of the ocean as the next frontier for economic development.

The EU’s digital twin of the ocean

Digital twins are built as replicas of large systems, originally closed engineering and production systems but recently also open-ended and complex natural systems such as the ocean, climate and the entire Earth system. Using past observations and continuous real-time data of present behaviors, they propose to model and forecast future scenarios and are meant to act in the physical world through guiding management. Main features of digital twins are connectivity and interactivity, in that they aim to supply citizens and decision-makers with up-to-date science-backed knowledge to protect or otherwise manage environments (Jones, 2020). The development of large-scale modeling of the ocean as a digital twin represents a new development in ocean governance, with increased reliance on digital infrastructure. The two-way exchange of information between marine ecosystems and the digital twin is meant to create a feedback loop between the digital and the physical realms, which is emblematic of how environing media operate, as explained in the previous section. This means that as the physical ocean is datafied, that is, increasingly subjected to the collection of data points, the digital twin can supposedly model and then implement different what-if scenarios and impacts of human-ocean dynamics, so that decision-makers are presented with a clearly delineated set of choices with preemptively identified consequences. Examples of what-if scenarios include ocean warming, changes to the seafloor, sea-level rise, climate change, and extreme weather events.

The EU’s digital twin of the ocean (DTO) is currently the most advanced digital ocean-initiative, repeatedly presented as the bedrock of the EU Mission to Restore our Ocean and Waters,’ and features as a key priority in high level initiatives like the UN ocean decade and the G7 Future of the Seas and Oceans. The DTO is touted as a coherent, high-resolution, multi-dimensional, multi-variable and near real-time representation of the ocean that integrates different new and existing data sources with modeling, artificial intelligence and high-performance computing. In reality it is not an entirely new creation, but consists of a platform that interconnects a number of existing digital infrastructures of delimited ocean sectors and areas and makes them interoperable; the DTO builds on existing European digital resources and combines and integrates data from the European Marine Observation and Data Network (EMODnet) and the Copernicus Marine Service into a single digital framework. The DTO will supposedly include both real-time updates and long- and short-term forecasts of anthropogenic impacts as well as biodiversity conservation strategies and active management of marine ecosystems. It will also use AI processes, or machine learning, as “toolboxes” that scientists, entrepreneurs, citizens and others can use to predict events and plan activities in the ocean, making the DTO a “place of digital co-creation” with “unlimited” uses.

When the DTO prototype was presented in June 2024, the EU commissioner Iliana Ivanova asked the audience to “imagine a future where the mysteries of the ocean are accessible without the need for submarines, where technology and innovation help us monitor and restore ocean health through real time insights and also proactive solutions.” She then claimed that this had now been achieved with the new DTO platform. According to the presentation, the DTO’s central feature of so-called “what-if scenario analysis” will offer the union unprecedented “science-driven decision making” with a new approach to ocean resource management, mitigation and adaptation. Through real-time and predictive insights into changes in the marine environment, policymakers and politicians are supposed to be able to better identify challenges and impacting factors in different domains of the ocean. Reports on the prototype also describe the project as a “game-changer” representing a “quantum leap” in the EU’s efforts to achieve sustainable seas “thanks to artificial intelligence and digital technologies” (Loctier, 2024).

These descriptions of the DTO echoes what appears as exaggerated beliefs in the ability of digital ocean twins in general, such as the claim that “Digital twins, a nascent yet potent computer technology, can substantially advance sustainable ocean management by mitigating overfishing and habitat degradation, modeling, and preventing marine pollution and supporting climate adaptation by safely assessing marine geoengineering alternatives,” expressed by Tzachor et al. (2023) in a scientific article that questions the maturity, but not the essential nature or capacity, of the data and technology needed to create a digital ocean twin. As we will see in the next section, the promises or stated abilities of the digital ocean twins poorly match the reality of available ocean data, or even, on a more fundamental level, what seems possible to “datafy” and represent in this way in the first place. This is an issue not only for complex and under-studied ocean processes but also, and not least, for the social and economic dimensions that the DTO also promises to represent and integrate into its ocean predictions.

The construction of digital twins form an integral part of the EU strategy of the European Green Deal with ambitious targets of climate neutrality by 2050. The union has identified these developments as landmark actions in the process of greening the economy, and maintains that they will enable the desired shift away from less sustainable practices of development. The EU Digital decade 2030 policy package describes an urgent need to foster economic growth, improve productivity and competitiveness, by “twinning the green transition with the digital transition” (European Union (EU), 2024). It seems clear that the political hopes and stakes for a universal solution to the complex problem of economic growth and environmental degradation are high, and that digitalization appears to embody this desire. The new class of models that they represent, it is claimed by the director of development (Bauer et al., 2021), will close the gap in our ability to look into the future, disregarding the fact that modeling an aspect of a physical system along with its economic variables, such as Integrated Assessment Models do, will inevitably be incomplete and regard chosen parameters of interest (Asefi-Najafabady et al., 2021).

We can see here a drift from climate models to ocean models, to ocean digital twins, where the “cultures of prediction” that have come to dominate climate science and policy are making headway also into the ocean realm. The complex ecosystem of the world ocean is mechanistically reduced to its ability to help cut carbon emissions and providing desired resources, all dressed up in secure decision-making due to better modeling. This faith in modeling is reflected in Kestutis Sadauskas, the deputy director-general for Maritime Affairs and Fisheries at the European Commission, who claims that “it will be cheaper to make the decisions, and to make more correct decisions with less mistakes if we try to model it before we go out in real life” (Loctier, 2024). This comment is made after enumerating the needs for continued ocean extraction and exploitation in a sponsored news article, where the commission appears to be building trust for its blue economy through the digital twin of the ocean. The tendency to reframe complex political problems related to climate and environment as technical issues which can be solved with powerful diagnostic tools is clear here, and as pointed out by Saltelli et al. (2024) in their discussion of digital twins of the Earth, “these models provide policymakers with convenient instruments for justification and control.” In the following section we will place the DTO in its historical context of ocean environing, to provide a framework for understanding the claims and desires associated with its sociotechnical imaginary.

Knowing the ocean

Collecting data from the global ocean is challenging. The vastness of the three-dimensional marine space means that any singular data point is surrounded by a wide unknown, while its inaccessibility makes sampling both technologically and economically demanding. As a result, ocean data collection was for a long time, and still in many ways remains, a painstakingly slow and highly expensive undertaking. As such, the ocean has for most of history remained at a distance from human reach and control, particularly compared to terrestrial ecosystems. However, with technological advances, the amount of ocean data that is recorded and made available has increased sharply in recent decades. The International Geophysical Year (IGY), carried out in 1957-1958, marks the beginning of this increase, with an emerging scientific understanding of planetary-scale dynamics that included efforts to gather data that enabled the creation of “synoptic geographies” of the global ocean (Lehman, 2020). As seen in studies of the IGY as well as subsequent similar initiatives, in the making of such planetary views the infrastructural and material conditions of possibility tend to get occluded (Helmreich, 2009) and the resulting notions of Earth as a system are not self-evident reflections of a pre-given reality but emerge through a set of geopolitical and technical relations, which rests on imperial and colonial power structures.

The endeavor to research the planet as an inter-connected system took another important step in the 1980s with the rise of Earth System science (ESS), leading to new ways of environing, studying and governing the Earth system on a planetary scale (Chakrabarty, 2019). The ocean and climate were recognized as key intertwined systems shaping the planetary environment, leading to novel efforts to record climate-relevant, global-scale ocean data, such as altimetry measured by satellites and temperature data reflecting global ocean heat content and circulation (Höhler, 2021; Lehman, 2021). Another step-change in ocean data collection occurred at the turn of the twenty-first century, with the development of autonomous and remote-sensing technologies. One central such technology is the neutrally buoyant float, developed already in the 1950s but only used on a small scale during the twentieth century. In the late 1990s, this technology was institutionalized as a core ocean observation technology through the so-called Argo program, as the technology became sufficiently reliant and cost-effective to allow for large-scale implementation, enabling continuous and near-real-time data. Core variables recorded by the Argo floats are ocean temperature and salinity, which provide crucial data for climate models and climate change predictions. The continuous stream of data provided through the Argo program is also a prerequisite for even conceiving of a tool such as the DTO. Digital ocean twins can be seen as the most recent development, even culmination, of the gradual process of wet environing unfolding since the mid-twentieth century, and even longer, most recently propelled by the build-up of autonomous sensor networks and sampling process (Lidström et al., 2022). The generated data feed into so-called ocean general circulation models that describe physical and thermodynamic processes, showing the ocean’s role in the climate system, under changing atmospheric conditions as well as in absorbing greenhouse gases and acting as a sink to human emissions. Without this kind of quantitative data digital ocean twins would not be conceivable.

The EU DTO can also be seen as emerging out of the broad field of ESS, which emerged in response to, or as part of, the Cold War context of the 1980s and included the rise of environmental and complex system sciences like cybernetics, building on the recognition that human activities exert a strong influence on the Earth’s chemical and physical dimensions. Through military patronage as funding for Earth sciences, geophysics, including physical oceanography, experienced unprecedented growth. While monitoring and surveying the global environment had become a geopolitically strategic imperative, it also provided a lot of the information that would go into ESS. The older paradigm of direct observation gave way to “field instrumentation, continuous and quantitative monitoring of multiple variables and numerical models” (Steffen et al., 2020: 54), that is, environmental mediation by means of digital computers. James Lovelock’s (1979) Gaia hypothesis—the idea that life on Earth as a whole has a self-regulating mechanism that strives to maintain homeostatic and thus optimal conditions for life itself—was one of the more important epistemic precursors to this field. The key insight of Lovelock was the cybernetic notion of feedback and control between integrated systems which came from information theory and computing, which in turn would form the conditions of possibility for the emergence of the global environment as an epistemic object.

NASA, who formed the Earth System Science Committee in the 1980’s, actively planned and promoted the science using techniques of corporate branding efforts to secure the enormous amounts of funding needed. The notion that the planet should primarily be understood as a system, had a lot to do with the political economy in the U.S. of the time, while now often appearing as self-evident (Barton, 2023). The combination of existing media infrastructure in terms of satellites and computers, and a historically specific economic and politic situation, thereby contributed to environing the planet to appear in its current shape and form as a system rather than a web of interconnecting and self-organizing structures with emergent behavior, which has as much scientific support. This notion also feeds into the conception of the ocean as a cybernetic system of information and control that humans, equipped with advanced digital technology, can and should manage, control and steer, now embodied in the DTO.

Aside from what appears as exaggerated claims regarding the transformative potential of digital ocean twins for ocean governance and sustainability, the very notion of a “digital twin” of a natural system as vast, complex and data-poor as the global ocean is also fundamentally misleading. While vastly more data are available at present than ever before, the entire ocean is far from datafied. This is especially true for the deeper parts of the ocean: the central Argo program for example mostly provides data only for the upper half of the ocean. Estimates suggest that only 5% of the deep seabed has been mapped in detail and deep-sea biodiversity remains poorly known, with new species routinely discovered. In short, “data pertaining to fundamental oceanic processes and phenomena are partial, underpinned by gaps in scientific knowledge spanning the physical, biological, and chemical oceanography sub-domains,” meaning that any digital ocean twin will be, to put it mildly, “imprecise” (Tzachor et al., 2023). As described by Westerlaken (2024) with regards to digital twins of biodiversity, they shape a particular understanding of its object of twinning that have large-scale impacts on environmental futures by furthering a reductionist and fragmented mode of understanding driven by quantification and algorithms, while erasing more complex and relational knowledge of living entities in ecosystems. Similarly, Saltelli et al. (2024) finds that digital twins of the Earth espouse a reductionist scientific logic and reinforce and economistic vision of the environment, which privileges the feasible (available data) over the needed (vulnerable groups or sensitive outcomes).

In any realistic sense, global ocean data is always going to be incomplete, with some dimensions, such as geophysical parameters, more easily represented through the kind of data required by a digital twin, while others, such as biodiversity in the deep ocean, may be left out of the picture presented by this novel technology. Or, as Sheila Jasanoff (2017) puts it, “the term data tends to sanitize the world of observation, erasing from view the observational standpoints and associated political choices that accompany any compilation of authoritative information.” The same is true of scientific and technological feasibility; suggesting that a digital twin of the ocean is possible is, on the most basic level, misrepresentative of what kinds of scientific knowledge of the ocean is available and demanding of significant caveats.

A sustainably developed ocean?

The epistemic change from an ocean of wilderness to a marine environment increasingly subject to measuring, monitoring and management has been accompanied by recognition of far-reaching human impacts on marine ecosystems; increased data collection has shown the extent and reach of anthropogenic changes including extensive warming, pollution, overfishing and biodiversity loss, leading to calls for marine protection and conservation. At the same time, the technological advances that have enabled scientific progress also support expanding ocean exploitation, at a growing scale, depth and distance from land. Helped by digital technologies and satellite data, commercial fishing for example has been able to increase catches despite continued collapses of stocks. Deep-seabed mining, perceived as a potential future industry since the 1960s, has become increasingly feasible with advances in deep-ocean technologies (Sparenberg, 2019), while measurements of ocean carbon uptake are used to promote ocean-based technologies for carbon dioxide removal and storage (Chai et al., 2020).

The tension between environmental protection and marine exploitation and its relationship to scientific knowledge is of course not new. From the vantage point of the twenty-first century, it seems clear that the postwar period saw a rapid increase in pressures on the Earth system linked to the technological development that made possible a hitherto unimaginable level of production, the benefits of which have been very unequally distributed. Originally published around 2004 and then updated in 2015, the graphs of what has been termed the Great Acceleration from 1950-2010 show the rapid increase in twelve essential human activities and the impact on corresponding aspects of the Earth system: population, Real GDP, primary energy use, water use, fertilizer consumption, transportation and telecommunications have all been on a very steep upward curve in these now iconic graphs. In the Earth system part, mirroring steep curves can be observed in CO2, surface temperature, ocean acidification, marine fish capture, domesticated land, tropical forest loss and terrestrial biosphere degradation. Already in the 1970s, ecologists recognized economic development as the primary driver of environmental degradation, which would later become understood as a destabilized Earth system leaving Holocene conditions for the Anthropocene. The trend continued, and in 2015 the UN created and unified around the Agenda 2030 framework, to be implemented through 17 Sustainable Development Goals (SDGs), including 169 targets and 232 indicators.

While sustainable development is a contested concept—at least in some academic circles—it has become the main framework and guiding principle for implementing political and policy efforts to protect the global environment. The SDGs are themselves the result of massive and often vague quantification of data and science that necessarily leaves out many aspects of the world, but as they are interpreted as inherently good—leading to sustainability as an objective and uncritical outcome, with equal benefits for all—they can easily be mistaken for the whole picture. Moreover, it is by now clear that few if any of the SDGs will be met and that, for the environment, the negative trend overwhelmingly continues. According to the 2023 Global sustainable development report (UN, 2023), in the current state of progress based on select targets only 9.c.1 “increase access to mobile networks” and 17.8.1 “increase Internet use” are making substantial progress and are on track. The effort to achieve the goal for the ocean in particular, or SDG14, has been described unceremoniously as “a round and inclusive failure” (Andriamahefazafy et al., 2022). The fundamental disconnect between the green and digital transition should be clear, and even more so as resource and energy use skyrocket with the latest AI boom pushing through evermore hyperscale datacenters around the world, all the while programmatic overshoot (Malm and Carton, 2024) has become the norm and the 1.5 C degree target is out of reach.

One central issue inherent to the SDGs is goal conflicts, which occur when the achievement of one goal or sub-target risks hindering that of another. For instance, there is a clear dissonance between the goals of sustained economic growth and environmental protection, and in this way, the SDGs risk “serving as a smokescreen for further environmental destruction” in the name of sustainable development (Zeng et al., 2020). Similar goal conflicts are at the heart of calls for further support for an ongoing blue economy which is creating a blue acceleration (Jouffray et al., 2020). This discrepancy is present in all the materials presenting the EU DTO, in statements such as that the digital twin will provide knowledge that “will help design the most effective ways to restore marine and coastal habitats, support a sustainable blue economy and mitigate and adapt to climate change,” in a rhetoric that promotes the assumption that these are compatible rather than conflicting aims.

To remedy conflicts between different aims, a critical view of how different areas, spheres and systems interact and how trade-offs between these interactions can be understood and managed is essential. Rather than recognizing conflicting aims, environmental technology development by corporations, states and intergovernmental actors is embedded in a discourse that suggests that they are able to advance both economic development and decrease environmental impacts while sustaining current modes of living in the name of sustainability. This idea is deeply rooted in the discourse on sustainable development and features prominently in the so-called Brundtland report, Our common future (1987) that introduces sustainable development as an intergovernmental aim for the world’s nations and which speaks of increasing the Earth’s carrying capacity through technological innovation. It continues in Agenda 2023 and is at the heart of the Integrated Assessment Models (IAMs) which inform most climate policy today, in applying a discount rate for future technologies to remove carbon instead of mitigating emissions in the present in order to maintain economic growth. It recurs in the discourse surrounding the digitalization of the ocean that coincides with a rapidly increasing need for better and more sustainable ocean management in the present, as well as new means of using marine resources for human needs.

The current hype surrounding digital ocean twins and its relationship to an emerging “blue economy,” powered by digital twins that should enable us to transcend or resolve the tension between exploitation and protection, needs to be understood against this background. The uncritical statements for how the EU DTO will contribute to sustainable marine development and SDG14 reflect none of the controversies or challenges increasingly attached to Agenda 2023 and the sustainable development concept. The history of an intertwined relationship between ocean data and marine exploitation clearly demonstrates that more data has not in itself lead to better governance (Bakker, 2022; Wickberg et al., 2024), an assumption that nevertheless permeates the discourse surrounding the DTO. The promise of a science-driven decision-making model made possible by the digital twin that will be able to overcome contested aims for environmental protection disregards deep-seated goal conflicts and competing interests between perceptions of the ocean as a frontier for resource exploitation (including new interest areas such as deep-seabed mining and exploitation of marine genetic resources, or MGR) and waste recipient (from “storage” of “excess carbon” captured from the atmosphere to plastic pollution) versus views that emphasize the need for reducing human impacts and allowing, or assisting, recovery of marine ecosystems. There is nothing that supports the inherent assumption in the promises made for the DTO that a technology like digital twins in itself will create the prosperous sustainable green or blue future that the vision of its proponents promises. These claims need to be contrasted with the reality of half a century of failure in meeting environmental targets and foster a sustainable economic growth that does not generate excessive economic externalities. A digital twin of the ocean in terms of technology may have the potential to help ocean policy and governance to be better informed in some ways, but the surrounding discourse reinforces a reductive view of the ocean aligned with the needs for economic growth. In reality, views of how much anthropogenic impact and change of the world ocean is desired or tolerable is a deeply political one and not an algorithmic problem.

The DTO promises to empower citizens, inform politicians and help businesses in their quest for a sustainable blue economy, transforming “data into knowledge for everyone’s benefit.” Modeling, however, is always the result of choices made by the modeler and the data that is available or chosen. As Tzachor et al. (2023) have pointed out, the DTO is a thoroughly “Euro-centric endeavor,” designed to support EU interests, using European datasets, and controlled by EU members and officials. The result is that the DTO will inevitably be biased in terms of the scientific knowledge, political interest and cultural perspectives that are built into and supported by the model. While it might not be surprising that an EU initiative serves EU interests, the control of management of the world ocean should not be reserved for actors with economic and geopolitical dominance. A related observation is made by Drakopulos et al. (2022), when they demonstrate how the digital platform of Global Fishing Watch that combats overfishing risks leaving the structural drivers of environmental change unaddressed as the maps and datasets are only legible through “hegemonic geopolitical and political-economic orders deeply implicated in industrial (over)fishing.”

Furthermore, while the digital twin technology may impact our understanding of and relationship with the ocean, as pledged by the EU, it is far from given how extensive or, as is routinely claimed, transformative that impact will be, and it is especially unproven that it will resolve any of the sustainability challenges and goal conflicts that surround the ocean, as is also often claimed. The stated vision of the EU is to empower a shared responsibility to “monitor and enhance” marine habitats (with unclarity surrounding the meaning of “enhance”) and at the same time support a growing blue economy, two supposedly combinable aims. The sociotechnical imaginary of the DTO thus presents us with a narrative that a balance between sustainable exploitation and conservation or protection can be had through the use of increasingly sophisticated digital technology. As an environing medium, the digital twin of the ocean increases and accelerates the reach of human influence of the ocean environment driven by what is essentially cybernetic fantasies of information and control of a closed system.

This narrative is attractive, but also fundamentally questionable. While digitalization may be leveraged for both increased and perhaps more effective marine protection, it may likewise be adopted for advancement of the extractivist paradigm seen at land (Sörlin, 2022), and it is not clear how it could do both at once, by mitigating negative effects of economic development, in ways made possible by this new technology specifically. Notwithstanding such doubts, the creation and implementation of a digital ocean twin presents politicians and policymakers with an attractive way of promising to overcome negative anthropogenic impacts through technical means.

Another important and often highlighted aspect of the DTO is the integration of AI into ocean governance, and one recent paper ties hopes of its capability to simulate human intelligence to solve difficult problems related to the implementation of a sustainable blue economy (Brönner et al., 2023). The DTO is claimed to be able to solve the complex task of “balancing economic, social, and environmental considerations to ensure the long-term health and productivity of ocean ecosystems and the communities that depend on them.” While AI systems and their integration in digital twins can certainly be helpful in automating mechanical tasks and visualizing big data, it is crucial to understand the limitations and risks involved in applying algorithms for vulnerable human and more-than-human domains. As a growing body of scholarship has demonstrated, AI systems has tended to increase inequality, racism, sexism and other structural divisions of our societies while accelerating environmental impacts (Chun, 2021; Crawford, 2021; Bender et al., 2021; Wickberg and Gärdebo, 2023).

The ocean under human control

A final point we want to bring up about digital ocean twins is the peculiar promise of a two-way traffic between model and ocean, as this ties in quite literally with the theory of environing media. The digital ocean is the result of massive while at the same time limited amounts of sensors in the physical ocean, while the digital twin should ideally mirror what the data and modeling tells us of this digitalized ocean—the twin can then test different impacts of different decisions so that decision-makers may tinker with biogeophysical impacts to reach a desired state of the ocean by artificial means. This is a curious case of how geoengineering is included in the background of this sociotechnical imaginary, but ultimately presents us with an ocean imagined to be fully under human control. The role of digital ocean twins in the contested development of potential marine geoengineering strategies is considerable, with the allure of “digital safe spaces” where proposed technologies could be tested as centrally informing and used to motivate resources for both types of endeavors (Tzachor et al., 2023). It is worth emphasizing in this context that what digital twins actually mirror is the digitalized ocean and its vast sensor networks rather than the whole physical ocean itself, meaning that what may appear as a digital safe space is likely to be more of a proxy, in the best case, than a true representation of the consequences of any substantial intervention on the marine system. The fantasy or illusion of control plays an important part in promises for sustainable development—by claiming to be able to predict and continuously monitor changes in the marine environment, industry and others seek to ensure decision-makers as well as the public that interventions they make in the name of the blue economy will be environmentally “safe.” This rhetorical use appears as one of the main drivers for the creation of digital ocean twins. By presenting a model that seemingly represents reality, the ocean is viewed as a system fully brought within the control of people and systems of economic activity, with environmental impacts immediately identified and addressable. The reality of human knowledge and control of the ocean realm remains far removed from this vision. In order to achieve a more truly sustainable relation to the ocean through technical mediation, a digital twin would need to abandon the technocratic “view from nowhere” (Haraway, 1988), which obscures how knowledge claims are always situated, and integrate and make visible the actual environmental conflicts and ecological costs that come with increased economic development at sea. It would also need to recognize that decisions about the ocean’s future are not only technical or scientific, but also philosophical, ethical, historical and above all political. The move toward a blue economy comes as land-based development is increasingly recognized as putting the biosphere under immense pressure resulting in mass extinction of species and land degradation on a level hardly imaginable some decades ago. Because the ocean remains comparatively less studied and developed, moving economic interests there appear as more sustainable than they really are, and the digital twin of the ocean plays a key role in advancing this narrative. The idea of an ocean so vast that it is resilient to extensive economic development is a stubborn and fundamentally misguided idea of the twentieth century that lingers on in many of the efforts to develop the marine realm. As is clear from an overwhelming amount of marine science however, the oceans are under immense and increasing anthropogenic pressures with far-reaching changes in terms of fish stocks, warming, acidification and deoxygenation. The implications of these changes for the planetary environment are so far unknown, and will not allow themselves to be predicted or managed through any digital twin, of the ocean or the Earth. Informed decisions should start from this conviction rather than outsourced algorithmic expertise in a digital twin.