AI in the Wild: Sustainability in the Age of Artificial Intelligence
by Peter Dauvergne
Published September 2020
The Truth about Nature: Environmentalism in the Era of Post-Truth Politics and Platform Capitalism
by Bram Büscher
University of California Press
Published December 2020
Program Earth: Environmental Sensing Technology and the Making of a Computational Planet
by Jennifer Gabrys
University of Minnesota Press
Published March 2016
Environmentalists fiercely argue over the role of technology in advancing sustainability. To some, technological progress embodies the root causes of contemporary environmental problems. To others, a retreat to premodern times is neither practical nor desirable. There are, of course, many other forms of environmentalists. But while positions regarding conventional technologies – like nuclear energy and genetic modification – are deeply encrusted, in terms of new technologies – like AI and robotics – the contours of the debate are still raw. Perhaps it is time, then, to ask: What can digital technologies offer the environmental movement?
Applied to the environment, digital technologies could be categorized in four types: technologies for data analytics (such as machine learning), technologies for sensing (such as remote sensing or camera traps), technologies for sharing information (such as through social media), and technologies for intervention (such as robotics or automated cars). All these technologies interact, and shape each other in ways that create both opportunities and challenges for collective action around the environment.
To understand how environmentalists can and should relate to digital technologies, I turn to three recent books. Peter Dauvergne, Bram Büscher, and Jennifer Gabrys each explore a different aspect of the so-called digital revolution and its application to environmental governance. Dauvergne’s book AI in the Wild examines the use of ‘Artificial Intelligence’ technologies for conserving and managing ecosystems. Büscher’s book The Truth about Nature considers the use of social media to mobilize environmental awareness and funding. Gabrys’s book Program Earth asks how advanced remote-sensing technologies are transforming environmental knowledge production and citizen participation. The authors share an overarching insight: each of these technologies is powerful, but they are more seductive, and even more destructive, than they are productive.
In synthesizing these books, I offer three main takeaways. First, new technological innovations have their own environmental impacts. Innovations like smart cars and smart cities can improve energy efficiency, but they also generate their own negative environmental externalities. For instance, these technologies require minerals like tantalum and rare earth elements; these must be mined, and they must create new forms of spatialized land degradation. Moreover, many of these devices last only a few years, often by design, with little recycling potential. As a result, landfills pile up even as, paradoxically, more companies embrace sustainability. Thus, digital technologies may appear immaterial, but they are not without their own serious environmental consequences.
Second, these technologies can further enmesh consumers and producers with the complex supply chains of excessive consumption and production – the very political economic arrangements that cause environmental problems. It is easy to be lured by the possibilities of enhanced efficiency. Consider, for example, smart meters that track and reduce our household energy use without us having to alter our behaviors, or automated cars that rely on electricity and improve fuel efficiency. Yet the history of technologically driven efficiency gains warns us – what we gain with improved efficiency, we typically lose (and more) with increased consumption.
Lastly, if the core challenge is political, and not technical efficiency, the true test of a tech’s greenness lies in how it inspires collective action. In this case, it is not simply about designing fancy technologies but about doing the hard work to integrate them with existing institutions like environmental law and social movements. But technologies like social media animate a type of social consciousness that may be too short-lived and too selective for a sustained environmental movement.
Still, when mobilized for long-term social and environmental goals, rather than to sell new green products, digital tech stands to offer much to the environmental movement.
Consider an example of using digital technologies for environmental goals specific to New York. Most New Yorkers may associate the Gowanus Canal with a long and difficult history of land degradation and sewage dumping. The canal has been a site of nonstop industrial pollution, since the mid-19th century. Despite good intentions, most initiatives to clean up the canal have ended in deferments, blame shifting, and failures. A particularly intractable problem is the discharge of toxic sewage from nearby industries, exacerbated during storm events.
Gowanus Canal Conservatory (GCC) is a community organization in Brooklyn, New York, devoted to caring for the Gowanus Canal. Recently, they partnered with Temboo, a technology start-up, to develop new ways to monitor the Canal’s water quality.
Temboo’s signature technology is a no-code platform that consolidates and processes urban environmental data. Temboo prides itself on its highly accessible products, community engagements, and sense of social justice. Posts on Temboo’s website spotlight community collaboration, environmental justice, ethics, and volunteerism, registering its willingness to lean into environmentalist narratives about collective action.
The Temboo-GCC partnership is meant to address this problem of toxic-sewage overflow during heavy rains. Working with volunteers, Temboo and the GCC set up sensors along the Gowanus watershed to monitor storm runoff. The sensors send data to devices in the volunteer’s homes, which is then transmitted to Temboo’s proprietary no-code platform. The platform is also linked to many other environmental datasets, including New York’s Street Tree map. The data-infrastructure provides a new way to quantify how trees mitigate water pollution in the Gowanus. This data is then used by GCC to solicit investments in urban green infrastructure and to spotlight the need for action.
The Temboo-GCC partnership illustrates the potential, the limits, and the performances of mobilizing digital technologies to pursuit of environmentalism. Indeed, the partnership has produced a new way to quantify the ecological benefits of urban tree planting. Sensors, remotely controlled robots, and satellites can collect data in places humans cannot safely visit. The capacity to link different datasets can also provide more timely, continuous, and comprehensive information.
Yet if the final takeaway from this elaborate exercise is to show that ‘trees matter’ to cities, it feels like much ado about nothing. New ways of collecting environmental data are important, but the real question is how data translates into collective action. The technologies do not, themselves, answer this question; rather, we see how the technologies are embedded in larger institutions including law, economy, and community.
This also requires some self-reflection on part of technologists. Many believe digital technologies have a revolutionary and democratizing potential, but much of that does not square with pervasive platform capitalism and the massive economic concentration we are witnessing today. So many technologies that profess to be an alternative to the state only deepen corporate power.
Instead, it is worth asking more specifically: How do digital technologies enable (or disable) pro-environmental collective actions? How do these forms of actions differ across environmental, spatial, and technological contexts? To what extent can we see such initiatives address the root causes of environmental degradation, which, in the case of Gowanus, is the complex history of industrial pollution?
The three books reviewed here offer some answers. Dauvergne’s AI in the Wild starts by acknowledging the scale and scope of environmental change. According to the author, the human race is hurtling toward a socio-ecological catastrophe (if it is not already in it), and mankind must consider all solutions, including technology.
In the author’s signature style, Dauvergne relies mostly on examples to develop this argument. A recurring example in the book is the RangerBot, a semiautonomous submersible used in coral-reef conservation in the Great Barrier Reef. The bot is unique in that it can reach deep into coral crevices to identify and eradicate coral-harming starfish. No human diver can reach these places without putting the corals at risk. But such innovations are few, according to Dauvergne. At best, such innovations may generate occasional successes, but the overall environmental gains will eventually rebound into more extraction, more consumption, and more production.
When sharing examples, Dauvergne displays genuine excitement and admiration for AI technologies. Even so, the book’s overall argument is about the need for caution and for more decisive state steering of technological solutions. Dauvergne concludes that AI technologies are only useful on the margins; they can make our economies more environmentally efficient. But they also create the grounds for expanding the economy. ‘Artificial intelligence, we need to keep in mind,’ concludes Dauvergne, ‘has no capacity to overthrow the entrenched interests that are exploiting people and nature.’
Pifworld is an example of a ‘do-good platform,’ according to Büscher’s The Truth about Nature, a social media platform where various social and environmental causes are presented. Users can engage them in many ways, including through donations, hence the Pifworld tagline: ‘Change the world – your way.’ One specific project on the website is to fund elephant-conservation efforts in Southern Africa. But when Büscher traces the project to Africa, he finds that the Pifworld funding had little to do with on-ground conservation programs. In fact, during interviews, Büscher was mistaken for a representative of Pifworld and excoriated by local officials.
Büscher’s Pifworld, then, highlights the divergence of how nature is represented selectively and spectacularly on social media, and the real-world challenges of coordinating multiple interests and trade-offs when implementing conservation initiatives on the ground. In paying virtually for an elephant corridor, narratives about race, colonialism, and fortress conservation are easily lost.
The Truth about Nature is thus critical in ways similar to AI in the Wild, but where Dauvergne focuses on physical technology, Büscher critiques the role of social media in promoting new forms of environmentalism. Social media platforms like Pifworld can expand awareness of environmental causes at a scale and pace unlike other media genres. It seems reasonable to ask, ‘Why not use this power of social media to mobilize collective action around the environment?’ Büscher calls this the politics of ‘platforms for good.’
While the role of media in shaping environmentalism is not new, social media is somewhat different. Older nature documentaries often represented nature as a spectacle; in David Attenborough’s documentaries, for instance, we see ‘spectacular natures’: beautiful vistas with never-before-seen footage of animals mating and combatting each other in the wild. By activating donations and tourism, these documentaries come to determine the priorities of environmentalism. They decide what kinds of nature to protect, who should do the protecting, and what is considered a threat. Social media versions of spectacular natures portray something similar and may not rely on Attenborough-sized budgets. On our phone screens, spectacular natures come alive with smaller budgets, one shaky video flowing seamlessly into another.
‘Platforms for good’ are problematic for multiple reasons, according to Büscher. First, this kind of environmentalism is at the whims of changing trends. Second, by making the platform essential for collectivization, it concentrates power in few platform companies. Third, by trying to make conservation ideas digestible for a broader public, important context is lost, with detrimental implications for advancing conservation goals on the ground.
Büscher offers many examples, and I recapped the one that struck me most – Pifworld. With this and other platforms, Büscher shows how local communities are rarely engaged, ideas of fortress conservation are reinforced, and complex social phenomena, like poaching, are represented too simply on do-good social media.
Projects like that of Temboo that use sensors to monitor urban environments not only transform the city infrastructure but also determine what it means to be a responsible urban citizen. This is the key insight in Gabrys’s ‘Program Earth.’ Gabrys uses a broad definition of sensing, analyzing hi-tech solutions from water-monitoring sensors and robotic dogs to lo-tech solutions like citizen science. Gabrys concludes that different ways of sensing natures produce different kinds of citizenship, and thus should be understood as political technologies.
The book makes a very important point about environmental data: ‘Measurements are articulations of relevance.’ For science, relevance is about precision or accuracy, but, for artists, it might be about different experiences, and for the community, it might be about sustaining relations with and in nature. It is, then, not only about what new measurements make evident but also what kinds of new experiences, subjectivities, and tensions the very act of measuring opens up.
One of my favorite discussions in the book is about citizen science and how it inspires new forms of social relations. Citizen science is the crowdsourcing of environmental information like bird sightings by amateur birdwatchers. There are far more amateur birdwatchers than there are bird scientists, and systematically collating their sighting data can help scientific work. Yes, issues of accuracy, reliability, and validity complicate the scientific use of such data, and there are many ways to make citizen science data more accurate. Nevertheless, if the goal is to enable collective action, there are good reasons why low-quality citizen-collected data may be superior to high-quality remotely gathered data. This tension in balancing citizen participation and technical accuracy is a recurring theme through Gabrys’s book.
All three books help characterize how and in what ways digital solutions are featuring in environmentalism and to what end. More data about environmental systems and models are certainly useful, but asking how the data animates new forms of collective action is critical. Resolving environmental conflicts requires complex technical and social negotiations, including a recognition of competing interests and power imbalances. Framing technologies as the key factor holding back a progressive environmental agenda dismisses this messy reality. From an environmental perspective, I worry that environmental data becomes the end in itself – not the means to an end.
A larger concern is that new data-driven innovations create an active economy around digital environmentalism. References to technology buzzwords and partnerships with respected community groups grant technology companies legitimacy and funding. Unlike environmental NGOs of the past, which were disrupters of progress, these new collaborations frame environmentalism in the banal aesthetics of a new phone app, no different from what you use to order food or a taxi. This is a good marketing strategy and could be effective in drawing more funding toward environmental causes, as might be the case with the Temboo-GCC partnership. It may also awaken a new kind of environmental citizen, but to sustain this new environmentalism will require more than developing cool technologies.
The problem in environmentalism is not money but the lack of collective will to confront polluters, at the individual, community, and national levels. This is not to say that technologies and collaborations like Temboo-GCC cannot be constructive. But to substantively shift our current economic structures requires more radical forms of collective action. Tying the future of environmentalism to big tech will present incremental and occasional progress, but I’m worried that these technologies will mostly just enrich a few tech companies, not the environment.
The Sydney Review of Books and Public Books have partnered to exchange an ongoing series of essays with international concerns. ‘The Planet Needs Collective Action – Not Tech’ by Ritwick Ghosh, was published by Public Books on 20 September 2021.