Sage Journals: Discover world-class research

Abstract

The corporate discourse on the circular economy holds that the growth of the electronics industry, driven by continuous innovation, does not imperil ecological sustainability. To achieve sustainable growth, its advocates propose optimizing recycling by means of artificial intelligence and sets of interrelated datacentric and algorithmic technologies. Drawing on critical data and algorithm studies, theories of waste, and empirical research, this paper investigates ecological ethics in the context of the datacentric and algorithmically mediated circular economy. It foregrounds the indeterminate and fickle material nature of waste as well as the uncertainties inherent in, and stemming from, datafication and computation. My question is: how do the rationalities, affordances, and dispositions of datacentric and algorithmic technologies perform and displace notions of corporate responsibility and transparency? In order to answer this question, I compare the smart circular economy to the informal recycling practices that it claims to replace, and I analyze relations between waste matter and data as well as distributions of agency. Specifically, I consider transitions and slippages between response-ability and responsibility. Conceptually, I bring process-relation or immanence-based philosophies such as Bergson's and Deleuze's into a debate about relations between waste matter and data and the ambition of algorithmic control over waste. My aim is not to demand heightened corporate responsibility enacted through control but to rethink responsibility in the smart circular economy along the lines of Amoore's cloud ethics to carve out a position of critique beyond either a deontological perspective that reinforces corporate agency or new-materialist denunciation of the concept.

Keywords

Circular economy e-waste computational modeling critical data and algorithm studies sustainability responsibility

Introduction

In 2019, approximately 53.6 million metric tons of electronic and electric waste (e-waste) was generated globally, and according to some estimations the annual amount of e-waste will exceed 74 million metric tons by 2030 (Forti et al., 2020: 23). Branded as a solution, the circular economy promises to “design out” waste and keep products and materials in circulation through optimized reverse logistics, or a closed-loop supply chain. According to the narrative of the electronics industry, this implies that continuous innovation and concomitant consumption—critically referred to as planned obsolescence—no longer pose a threat to ecological survival. For instance, “Mining less from the earth and more from old devices” summarizes Apple's vision for a circular economy that supports production through recycling (Apple, 2018: 22). Against the de-growth mantra of environmentalists, the corporate discourse on the circular economy holds that growth of markets and continuous product renewal go hand in hand with sustainability (Hogan, 2018), as long as there are ways to optimize recycling. To develop these and “close the loop” an increasing number of corporations as well as governance actors seek solutions in artificial intelligence (AI) and interrelated applications of Big Data and Internet of Things (Dauvergne, 2020).

Recycling industries and waste management predate the deployment of datacentric and algorithmic technologies. Yet “smart” technologies seem to evoke a forceful imaginary of circularity that is shared by various, new and old, actors. Whereas previously considered a business that is labor-intensive, logistically complex, and rife with reputational and legal risk, waste management is now framed as an opportunity by big-brand technology manufacturers, governments, and investors alike (UNEP FI, 2020). One of the primary advocates of the smart circular economy is the Ellen MacArthur Foundation (EMAF), a globally leading charity that is affiliated with powerful tech corporations such as Google, Cisco, Apple, IBM as well as institutions such as the World Economic Forum and the consultancy firm McKinsey. Other interested governance institutions are the United Nations, International Telecommunication Union, and states, including China, which has historically absorbed much of the global e-waste through its informal sector (WEF, 2018). The founding of the International Standardization Organization Technical Committee 323 (ISO/TC 323) for Circular Economy in 2018 further indicates that the circular economy has become well-established as a policy framework (Xavier et al., 2021).

According to EMAF, there are three areas to which AI can be applied: (a) product design and testing; (b) circular business models and predictive analytics; and, central to my study, (b) reverse logistics and optimization (EMAF, 2019: 4–5, 9–10). Ubiquitous datafication and surveillance are commended as a means to undercut not just waste itself but informal recycling. As EMAF formulates it, while existing systems and processes of recycling are “cumbersome, fragmented and labour intensive” (EMAF, 2019: 30), a staggering “80% of electronic waste is not treated appropriately as a result of poor collection, technical complexity and the cost of recycling, remanufacturing and refurbishment” (ibid.). Citing a similar number that was calculated by subtraction of collected waste volumes from sales volumes, the Global E-waste Monitor (Forti et al., 2020) reports that, “globally, only 17.4% of e-waste is documented to be formally collected and recycled” (p. 23), while the rest disappears either in dumps and landfills or enters informal recycling circuits. Pointing out malpractice and disorder, the smart circular economy's stated goals of sustainability and corporate responsibility come together with a bid for corporate agency and control.

It should be noted that in practice, the formal and informal circuits currently often intersect (Bisschop, 2012; Gabrys, 2013) and technologies of surveillance can turn into technologies of neglect that aid informality, non-governance, and illegal practice (Rossiter, 2016; Tsing, 2009). As I (Hoyng, 2019) have argued elsewhere by drawing on the notion of the technological “gaze,” rather than working toward all-encompassing datafication and omniscient vision, the infrastructural apparatuses of recycling both turn a “blind eye” toward informal settings in which regulatory violations may take place and they cultivate liminal “blind spots” within their datafied constructions of efficiency. Further technologization of the circular economy through blockchain applications promises a fully transparent supply chain, but in practice this seems neither very feasible nor desirable for actors along the (reverse) supply chain, including the transnational corporations that rely on outsourcing (Posner, 2021). An important line of critique of waste and recycling practices addresses the lack of corporate integrity and their toleration of environmental violations. My argument, however, takes a different route. My aim is to consider and critique the smart circular economy on its own terms, as envisioned in designs and concomitant discourses that seek to signify and domesticate the potential of new datacentric and algorithmic technologies in waste mining and logistics. I focus on emergent circular economy apparatuses and imaginations (Friant et al., 2020): technologies that track, surveil, index, calculate, and model e-waste, along with datacentric rhetoric and aesthetic of transparency in glossy, lucid-looking corporate environmental responsibility reports, featuring lists, charts, tables and infographics. My question is the following: how do the rationalities, affordances, and dispositions of datacentric and algorithmic technologies perform and displace notions of corporate responsibility and transparency? In order to answer this question, I compare the smart circular economy to the informal recycling practices that it claims to replace. I analyze relations between data and matter as well as distributions of agency inherent in, respectively, the informal circuit and the smart circular economy. Specifically, I consider transitions and slippages between responsibility and response-ability. Conceptually, I bring process-relation or immanence-based philosophies such as Bergson's and Deleuze's into a debate about relations between waste matter and data and the ambition of algorithmic control over waste. I approach the question of responsibility in the circular economy, as it is envisioned and promoted in times of datafication and automation, by comparing and contrasting formal transparency/control with informal opacity/openness as well as formal and informal (ir)responsibility.

This study makes use of several reports from the EMAF, the World Economic Forum (WEF), the United Nations, the International Telecommunication Union, and the European Environmental Agency (EEA) belonging to the European Union. In addition, I consult scientific research papers from fields such as computer engineering, which propose algorithmic and datacentric applications to “close the loop” on products and materials. Especially, several literature reviews that deploy databases such as Scopus, Web of Science, and Google Scholar provide a useful overview of the scope of the current technological imagination (Banasik et al., 2018; Nobre and Tavares, 2020a, 2020b; Pagoropoulosa et al., 2017; Paul et al., 2021). In my analysis, these resources complement each other: the scientific papers unfold a technological imagination that is rather specific. They allow me to break down the broad area of algorithmic technology and understand specific technological rationalities and affordances of design. The celebratory advocacy texts by EMAF show the cultural metaphors and discourses that construe technological operations in relation to organizational logics of the circular economy. These metaphors and discourses are performed in practice, and hence have material effects that require further scrutiny (Agostinho, 2019). The governance reports reveal discursive struggles and shifts related to transparency and responsibility at times when datacentric and algorithmic technologies inform ecological ethics. Furthermore, to compare and contrast formal operations to informal practices of reuse and recycling, I conducted fieldwork at markets in Hong Kong that house the informal sector. Drawing on visits to the markets in Sham Shui Po district and the trade and repair shops in Chungking Mansions and Sin Tat Plaza, I study display methods, complemented by brief explanatory interviews with repairers, sellers and customers. This gives me insight into practices of reuse and conceptions of product and waste in the informal recycling sector, especially taking cues from the ways in which items are circulated, stored, classified, and ordered.

The next section conceptualizes waste and its specific relation to datafication and computation. The following sections compare the informal and formal recycling circuits in terms of relations between data and matter and distributions of agency. Finally, I underscore the persistence of uncertainty in the datacentric and algorithmically mediated circular economy, which problematizes corporate notions of responsibility and transparency.

Waste's potential and liminality

Process-relation or immanence-based philosophies such as Bergson's and Deleuze's (Ivakhiv, 2014) underscore waste matter's virtual potential to “become.” Such potential is incorporeal and yet immanent to matter, forming a latent creative force that can destabilize fixed forms in processes of transformation. Arguing that this material potential is best exemplified by metal, Deleuze and Guattari (2005) write that metal's deceptively solid appearance conceals the variability of alloys that allows them to change and recombine. They note: “Matter and form have never seemed more rigid than in metallurgy; yet the succession of forms tends to be replaced by the forms of a continuous development, and the variability of matters tends to be replaced by the matter of a continuous variation” (p. 411). Whereas the blacksmith seeks to wield and provoke such potential skillfully, this kind of endeavor proves more difficult in other cases such as plastic pollution drifting through oceans. In this regard, Gabrys (2016a) distinguishes between the “object-ness—and the potential object-ness—of a garbage patch,” the latter being “never stable or given but in process and giving rise to new engagements.” In the case of e-waste matter (by which I mean components and materials belonging to electronic objects), both metals and plastics are at stake, among others. Virtual potential exceeds the state of discrete electronic objects and classified items, labeled by models and brands. It often starts to matter (literally) when disposed objects no longer carry out designated functions and lose shape and identity. Furthermore, waste matter's potential, its proclivity and predisposition for transformation, becomes actualized through particular material and social relations. Hence, material agency can manifest in ways technological (as refurbishment), chemical (as retrieved material such as plastics, metals, or rare earths), and ecological (as pollution). For instance, during refurbishment, material agency comprises the ability of solder to respond to heat in a capillary manner and fix or dissolve the connection between a charging port or battery terminal and circuit board. Regarding pollution, cadmium is a metal in rechargeable batteries and exposure to it has a slow, long-term impact on bones in human bodies. Brominated flame retardants in plastics, damage neurobehavioral functions, reproductive health, and the thyroid, depending on the recycling method.

The virtual potential and relationality of waste matter might explain why waste is often positioned as the liminal outside of knowledge systems, as the classical anthropological definition has it (Douglas, 1984). Waste as unstable material mass or flow cannot be properly mapped, charted, or broken down in countable units. Waste is what resists being turned into data and definition. But, also, what lacks data goes to waste, having no social utility (Offenhuber, 2017; Parikka, 2015). Arguably, waste matter's indeterminacy continues to complicate endeavors to know the matter even in the era of datafication and computation, reinforcing its position as liminal outside. Fazi's (2019) distinction between the virtual and the discrete, which builds on Deleuze's Bergsonian conception, applies: the virtual encompasses what I referred to as the potential associated with the “metaphysical plane of transformation of what is prior to individuation” (p. 5), or actualization. The virtual is not logically predetermined, but creates through difference, producing an “infinity of variations, modulations and transformations” (Fazi, 2019: 5). In contrast, the discrete involves “digital technology's processes of quantification and codification” that can merely “approximate, but never fully grasp, life's ontogenetic conditions” (ibid.). Computation forms “a numeric way of arraying alternative states” (Massumi quoted in Fazi, 2019: 10) that systematizes and calculates possibility and probability, but this does not equal the “pure potentiality in becoming” (ibid.). Hence waste as an embodiment of indeterminate virtual potential can be considered not just the liminal outside of knowledge systems in general but also, in particular, both the outside and conceptual opposite of the discrete modes of datafication, quantification, and codification at the heart of computing.

Informality

As my fieldwork at markets in Hong Kong has taught me, the informal circuit is characterized by the lack of information about objects and by opacity regarding practices and processes. Hence, I define informality as practices of repair, reuse and recycling that circumvent inspection and control by big-brand tech corporations and authorities, while, at times, defying hegemonic norms of consumer culture as well as environmental and other types of regulation. Exactly because the informal circuit of e-waste recycling applies data and information less strictly, it also loosens checks on the indeterminate potentiality of waste, which resonates with and through its modes of operation. Deleuze and Guattari's figure of the metallurgist, adapted by Parikka (2015) to characterize craftwork enlisting technomateriality, speaks to the interaction between e-waste matter and humans at informal sites of scavenging and salvaging. The metallurgist “follows the flow of matter,” which is “teeming with material potentials that can be teased out in different constellations, temperatures, and conditions” (Parikka, 2015: 23). A more empirical look into the informal sector would teach us that the figure of the metallurgist only grasps a certain dimension of the agency and labor involved (Isenhour and Reno, 2019; Jackson, 2014; Jackson et al., 2012; Offenhuber, 2017). In my own fieldwork observations, repairers often considered their own jobs as routine-based rather than exploratory or inventive, and the availability of no-brand spare parts standardized procedures. Also, other studies of clustering in e-waste salvaging indicate that informal sites can exhibit characteristics of formal industrial clusters (Davis et al., 2019), such as divisions of labor. Nonetheless, it is clear that the informal circuit facilitates different forms and degrees of involvement with matter and its potential. A telling example comes from an African second-hand electronics trader who had come to Hong Kong to buy up disposed devices in bulk and who told me that when he was a young boy, his dad had once smashed a television right in front of him. By demonstratively destroying the desired commodity, the father intended to teach his son how to look beyond the commodity as an integral object, engage technomateriality creatively and learn to fix devices (fieldnotes, Sham Shui Po, 25 February 2019).

The etymology of “informal” goes back to the mid-15th century and stipulates a condition that is “lacking form; not in accordance with the rules of formal logic” as well as “irregular, unofficial, not according to rule or custom.”¹ This etymology resonates with the unregulated, opaque, and often illegitimate practices of the informal sector. Its operations take place without too much data, sorting, and classification: there is unsorted or only roughly sorted stock and lack of detailed inventory. The lack of data that in-form (modulate and control) an “order of things” is reflected in display methods in the informal circuits of street markets: a heap of unsorted, hardly extractable cables, different devices or brands, or even electronic items mixed up with magazines, broken toys, and watches. Such display methods reflect opacity as much as an indeterminate ontology. Sometimes, items go in and out of bins as collectors and hawkers pick them up and dispose of them elsewhere, meanwhile imagining and speculating what they can do with them, if they can use any part of it, or if the item could spark desire for somebody else. Value (use and exchange) can be found in the minimal acts of relocating, waiting, and storing in the process of which waste might transition into something else through usage, or the prospect of usage, by somebody.

The loose categorization and labeling of matter, if any, creates opacity but also accommodates flexible status and relations for objects (Hoyng, 2018). Crucially, the boundary between “working” and “not-working” is flexible and reversible. Bulk sales often mix working devices and not-working ones, and buyers do not look at the state of devices they purchase in container-sized loads. Whether devices are still functional is often not a concern as the goods are shipped to places where local expertise in repair and refurbishment is abundant. Traders who buy a few hundred mobile phones negotiate a lower price, after estimating the state of the devices, and individual customers look at items with an eye toward how they could be fixed, refurbished, or recombined. For instance, a construction worker explained to me that he purposefully bought a broken phone on the Sham Shui Po market. The speaker did not work anymore, so he got earplugs in addition. According to him, he had purchased the two items for a bargain price and together they constituted a “functional” device, meaning one that served its purpose just fine. Yet pointing to the instability of technomaterialities in the informal circuit, he also quoted a Chinese proverb: “Expect the unexpected!” (fieldnotes, Sham Shui Po, 25 September 2016).

Along with “working” and “not-working,” between “real” and “fake,” there are many cultural categories for refurbished phones, 14-day phones, China-brand phones et cetera that are absent in hegemonic consumer culture. Refurbished phones may receive new covers and even boxes in order to provide a “brand new” and standardized look, while the inside of such phones can be composed of a mixture of components retrieved from other phones or non-original spare parts. Yet according to sellers, the outlook is not deceiving customers and the latter are by no means cheated. They supposedly knowingly buy a newly looking, yet refurbished phone. The designation of “fake” only pertains to cases when “cheaters” sell refurbished devices as new, original ones, which ultimately is indicated by the price. Sellers and traders perceive this as an immoral, if not criminal, act. Their judgment seems in line with the rationale of police interventions at markets such as Sin Tat Plaza, which target the obvious “cheaters” but not everyone. At street markets, there is likewise little concern about intellectual property violations or corporate norms advanced by tech brands that suppress refurbishment, third-party repair, and do it yourself (DIY). Hawkers are primarily afraid of officers of the Food and Hygiene Department, who patrol the streets to maintain public order, while the asylum seekers who are involved are afraid of ID checks by police, as they are not allowed to work while their cases are being processed.

In line with what Tsing (2015) designates as the nonscalable practice of salvage, value retrieval results from the coalescence of material potential and situated needs and desires; responsiveness to incidental occurrences; and adaptive tactics and tacit knowledge. The informal circuit leaves room for a variegated set of actors to respond to matter by means of situated knowledge, tactical intervention, and tacit skills. The opacity of informality combines with openness in the material relations between humans and technology, enabling multiple engagements, including tinkering, refurbishment or piracy. Hence, in the informal circuit, scavengers, traders, repairers, and users transgress the binary of consumption and production when they salvage materials at scrap sites or street markets (Bell, 2019). Sites such as Sham Shui Po in Hong Kong and Huaqiangbei in Shenzhen, situated across the border yet connected as informal zones (at least prior to the COVID-19 pandemic), cultivate a DIY culture. In fact, customers cannot remain just passive or irresponsive in front of a heap of objects or a pile of loosely sorted devices; matter demands to be pried and gauged. Many of my informants referred to embodied modes of knowing, touch, and sense. For instance, a repairer in Sham Shui Po considered his job a “craft.” Yet rather than the course he attended in Mainland China, he claimed that a repairer needs to have “talent,” consisting in a certain kind of sensitivity or intuition and steady hands (fieldnotes, Sham Shui Po, 18 May 2018).

In sum, the informal circuit enables response-ability as the agency and ability to respond to the specific, concrete presence of matter (Turner and Tam, 2022). Often, such response-ability implies the ability to be affected by matter—to be “touched” by it as much as to touch it—constituting an ethics of care toward matter. Whereas only some in the informal circuit indicated to be motivated by environmentalism, many reflected some kind of care for technomaterialities and talked about recovery in moral terms. Puig de la Bellacasa (2017) has argued that “to care” collides the meanings of maintenance and becoming affected, while touch enables a “situated response” to concrete situations that can be speculative in the sense that it gauges alternative possibilities in more-than-human relations. To be close to matter, as in the case of maintenance and repair, can involve love for a world of things and an ethics of care and responsibility vis-à-vis them.

At the same time, however, the openness of materiality that informality affords is not risk-free and in several respects the sector relinquishes moral consideration for profits. In this regard, the responsibility to figure out if something is working or not, “real” or “fake,” easy to breakdown as well as the risk of bad surprises, are distributed to customers at informal markets, along with the agency of response-ability. For customers who feel inapt at navigating the haptic landscapes of the markets, there are guides, such as the Ethiopian man living in Hong Kong who takes mainly African traders and individual buyers to second-hand shops (fieldnotes, Sham Shui Po, 28 May 2019). Indicating a widespread sense of risk, repairers complain about being mistrusted by customers. One of them negotiated the situation by putting up a camera above his workbench, connected to a big monitor facing the customers, to render the repair process witnessable and, thereby, “transparent” (fieldnotes, Sham Shui Po, 17 May 2018). But at many levels, operations in the informal circuit are opaque to anyone but those at the heart of the business. Along with subverting binaries of product-waste and producer-user, the informal circuit cultivates ambiguity around what constitutes legal trade versus criminal activity. A report by Interpol and EU explains that second-hand goods, waste, and repairable or reusable materials resist classification and unequivocal legal regulation (Huisman et al., 2015). Emphasizing the labor costs of testing and sorting, the report states that shipments often include a mix of things, such as still functional but “very old UEEE [Used Electrical and Electronic Equipment] with no real value or market anymore, or with very short remaining lifespans” (ibid.) as well as repairable waste electrical and electronic equipment (WEEE), and “relatively new but non-functioning appliances ideal for harvesting of spare parts” (ibid.). Typically, items originally discarded as WEEE change their status, meaning that at some later point these same items are no longer presented as waste so that laws restricting waste export/import do not apply. Evading or ignoring laws and standards related to e-waste trade, intellectual property as well as labor and environment, the informal sector—exceptions and personal concerns granted—relinquishes the ethical and/or legal question of responsibility.

Corporate response-ability

Remarkably, whereas the informal circuit avoids data and information, or applies it in flexible rather than a standardizing manner, the formal circular economy promises to optimize recovery and to exploit to the fullest material potential by intensifying datafication. In this way, it seeks to turn a problem—waste as unknowable, unmanageable, and useless matter —into an advantage: waste casts a specter of profit exactly because its initial lack of function and value makes it lucrative to mine waste (Gabrys, 2013). AI and other algorithmic technologies are presented as innovative means to probe the potential of waste matter and to operate at the threshold of knowledge where this was previously considered impossible by automated means. Current applications follow in a line of cybernetic and autopoietic systems that turn “the notion of the limit away from ideas of collapse” (Parisi, 2013: 16). The limit becomes productive, opening up new prospects for tactical intervention. For instance, deploying Big Data to discover new sources of value and function in waste matter requires finding ways to salvage “exhaust data” that initially appear useless (Thylstrup, 2019) and utilize messy, heterogeneous datasets to gauge waste's potential. Developing new ways of producing and “recycling” data to produce knowledge is thus a precondition for salvaging waste itself, challenged by the fact that data is often structured according to the logics of the linear rather than circular economy (EEA, 2017: 31).

As envisioned and emergent in practice, the formal smart circular economy relies on an amalgamation of interrelating technologies that facilitate everything from tracking to computational modeling in support of decision-making. Among the technologies deployed in recycling processes are, for instance, Radio Frequency Identification (RFID) tags, Near Field Communication (NFC) sensors, and Global Positioning System (GPS) navigation, which generate real-time data and help track goods throughout the reverse supply chain. Furthermore, automated recognition using image-based machine learning supports sorting e-waste types, mobile phones, and batteries classification, in addition to assessing the condition of electronic devices in a process referred to as “grading.” Featured as one of the success stories, EMAF (2019) reports that the company Teleplan (now Reconext) deploys AI “to help provide aftermarket and lifecycle care services to keep electronics products in use, such as managing reverse logistics, screening and testing, and repairing and refurbishing of used electronic devices” (p. 30). Machine learning is deployed to grade used devices supposedly in “objective, consistent and cost effective” manner by assessing the “external or cosmetic condition of the devices and determining if they can be reused or resold, whether they need to be repaired, refurbished or recycled, and their market value” (ibid.). Similarly, by using AI the company Refind enables other companies to “extract the full value from (mixed) e-waste streams” (p. 32). It uses AI for sorting by combining sensors and cameras with supervised machine learning for image recognition. Manually labeled digital images of objects are inputted and coded so that patterns can be systematically compared, and identities and conditions inferred in probabilistic manner. The claim is that the sorting systems “can classify the type, and if perceptible the condition of e-waste at a granular level” (ibid.). Such granularity in sorting and grading ideally enables “cascading” (EMAF, 2018) meaning that an item moves from reuse in high-end consumer electronics to lower performance applications and eventual material recovery. The intention is to retain the value and utility of products and components longer. As cascading constitutes an iterative act of filtering, sorting and categorizing, the virtual potential of waste matter is continuously probed and actualized in a series of concrete possibilities that changes over the multiple courses of reuse and recycling (Hoyng, 2022).

The operations of the circular economy introduce a practice of “making” in the face of the uncertain and the unknown, namely waste as liminal object carrying indeterminate potential that can be actualized variably, depending on the relations it enters into. Such making signifies a form of control: it registers, monitors, and secures all waste flows in such a way that data subsumes matter and reduces it to code to shape it and keep material agency and potentiality in check. Control operates as a formative and anti-entropic type of power that renders matter a sculpted materiality “in which vital agents are managed, organized, affected” (Galloway, 2004: 110). It draws on a second, mediated mode of response-ability, enabled by constant data input, ubiquitous computing and self-learning systems. As Massumi (2015, 123–124) writes, control is a mode of power that revolves around the ability to respond to emergent, imminent or possible events; hence, a responsiveness in support of interventions that modulate, stir and constrain developments in its environment. It requires decentralized “sensing” and situated, tactical decision-making. Likewise, following an expert cited by WEF, circular economies require decentralized and adaptive intelligence: “Truly circular economies arguably cannot exist without the Internet of Things” (Tim Brown quoted in WEF, 2015: 7). The quote continues: “No amount of clever design ensures a complex system will remain useful and efficient over time. To be sustainable, a system must be responsive; actions and behaviours must be connected via data and knowledge. With the embedding of intelligence in almost every object, we can imagine systems that adapt and respond to change in order to remain fit for purpose” (Tim Brown quoted in WEF, 2015: 7). In such statements, “sensing” via the Internet of Things appears to mimic decentralized and situated response-ability in the informal circuit, however with the aim of enhancing coordination and control in a way that actually prevents local actors from making decisions, which is seen as a risk and potential liability. Another way in which the smart circular economy emulates certain qualities of the informal circuit's response-ability involves its dealing with concrete singularities in automated manner. For instance, with regard to assessing material potential, the director of Teleplan (cited in EMAF, 2019) explains the challenge of creating a technology “that could not only ignore standard elements but also identify defects that have – by nature – no patterns, given no two defects are ever the same” (p. 31). Breakdown, in all its uniqueness, “was a mathematical challenge that only machine learning could solve” (ibid.) This statement echoes Steve Jackson's (2014) insight that breakdown and repair are sites of difference and variegation, but it argues that machine learning automates the processing of singularities without resorting to human sensing and touch, hence enabling the circular economy to scale up without accruing excessive labor costs.

Moreover, profuse surveillance and control of matter do not simply enhance reacting to emerging events. Rather, they afford envisioning presents and futures, which in some ways exceeds the abilities of actors in the informal circuit who respond and act on the basis of situatedness. Computational modeling in the service of decision-making produces predictive statistics and forecasts of “what could happen” as well as prescriptive recommendations, roughly including sales in aftermarkets, repair, refurbishment, remanufacturing, or recycling (Banasik et al., 2018; Paul et al., 2021; Tsoulfas et al., 2009; Zhang et al., 2019; Zhang et al., 2021). Such prediction/prescription translates waste's indeterminate potential into calculations of value that consider the probable and possible futures stemming from shifting consumer demands, costs, and supply chain needs, shortages, etc. (EMAF, 2019). As multiple futures are drawn into the present, waste is always already becoming something else and product-to-be.

It is by means of control as a formative force and anti-entropic type of power operating through “smart,” mediated response-ability that the formal circular economy performs corporate responsibility and distinguishes itself from the informal circuit. The technological infrastructure of the smart circular economy promises corporate responsibility: it corroborates oversight, self-reflexivity, and self-governance, enabling a slippage between response-ability and responsibility. Moreover, corporate environmental responsibility reports mobilize a rhetoric and aesthetic of transparency, supposedly rendering information about recycling operations available for scrutiny and no longer opaque.

The ontology of delimitation

Data and information in the formal sector do not simply play an instrumental role in optimizing material recovery but mediate the making of a world (Ruppert, 2012). We can turn to etymology once more to understand how. Deriving from Old French, “formal” as an adjective goes back to the late fourteenth century and refers to “form or arrangement” and also the “essence of a thing.”² Considered in the context of e-waste recycling, this etymology speaks to the ways in which the formal sector aims to uphold some kind of order and give shape to matter. Formal industries are in the business of in-forming matter on the basis of an assumed “ontology of delimitation,” which delimits the life of matter to a specific material order. For instance, cascading is not an open-ended exploration of matter's potential but a trajectory of pre-set alternative options, limited by the ontology of delimitation that identifies and indexes objects as particular things. Guarding this ontology of delimitation, categories including item, brand, series’ model and condition scores affix matter and constrain the material world of waste in the formal circuit, even as they also enable reuse and recycling. Indeed, sorting often upholds brands as a category to identify waste matter and authenticate provenance, as brands are key to aftermarkets, refurbishment and reuse of components. Whereas they are imitated and faked in informal circuits, the website of Reconext proudly announces that their automated systems are able “to inspect devices to identify fraudulent claims or counterfeit parts.”³ Throughout the formal recycling sector, whenever identification and sorting by brands cannot be maintained in practice, shredding is a very common strategy that produces clean streams of raw material (for instance, aluminum, gold, rare earths, copper, palladium, and plastics) that are easy to feed into formal supply chains of production, while avoiding quality, security, or rights complications (Laser and Stowell, 2020). Yet shredding annihilates reuse possibilities of objects and components and undermines the principle behind cascading to maximally exploit waste matter. It shows the tension in the formal circular economy between (a) exploitation of waste matter's transformative potential and (b) protection of the corporate ontology of delimitation. To the extent that the ontology of delimitation guides the operations of the datacentric and algorithmically mediated circular economy, it aids what Gabrys (2016b: 187), following Barad, refers to as thingification: “the turning of relations into ‘things,’ ‘entities,’ ‘relata.’” Thingification erases how things emerge and come into being, or how they could be different, as the underlying potentiality remains suppressed, constrained and kept in check. Meanwhile, contrary to the suggestion of perfect circularity, some materials simply remain waste, meaning that they are beyond recuperation, while the recycling and remanufacturing process requires the expenditure of further resources (Gabrys, 2013; Gille, 2010)

Moreover, data and information promise corporate transparency. But such transparency exists alongside the black-boxing of technological hardware, which users have no right to repair. Whereas the formal sector offers corporate transparency in environmental responsibility reports characterized by datacentric aesthetic, the technomaterialities of which electronics are composed, such as batteries, motherboards, RAM modules and processors, form a black-boxed constellation, concealed from view and literally sealed off with a warranty seal that inhibits unauthorized engagements. Meanwhile, control in the smart circular economy introduces a certain distribution of agency. Whereas users may be asked to participate in processes of production/disposal, they in fact have little agency. Prosumerist relations of interaction exploit and co-opt user participation. Especially in the case of the big brands that aspire to “close the loop” of production and disposal on an individual basis, users are involved in the logistical work of recycling through return-and take-back programs. For instance, users generate valuable data by sticking barcodes on return items or scanning QR codes. The data produced mostly unknowingly throughout the journey of the item in and out of the hands of the user culminate in sizable datasets, whose usability extends beyond coordination of the travel of the individual item back to the corporation.⁴ The interactive interfaces of reverse logistics extend production into not just the phase of consumption (Terranova, 2004) but of disposal. Brands enhance these prosumerist interfaces as they elicit consumer participation along with trust in the corporation (Lury, 2009).

In the smart circular economy, the corporation is the privileged agent. Hence the formal and informal sectors can be contrasted in terms of their respective performances of transparency paired with exclusive control and opacity paired with risky openness. Besides cultivating fixed producer and consumer identities, the smart circular economy actualizes waste matter's potential to “become” in particular ways by restoring definite form and identity, while foreclosing other relations and possibilities. These alternate possibilities do not only include illicitness in informal settings but also, more broadly, all those possible destinies and futures deemed less than “optimal,” as the next section will argue.

Ecological ethics in times of algorithmic computation

To further my critique of responsibility as response-ability, I would like to focus on the latest proposals for decision-making and optimization of waste mining that involve computational modeling. According to experts, computational modeling enables “detailed tactical and operational decision making” in reverse logistics, even when many factors are at stake and conditions are volatile (Goodall et al., 2019: 49). It assesses “benefits and risks of strategies […] faced with high levels of uncertainty and complexity” (Goodall et al., 2019: 49). To optimize outcomes, such modeling requires integrating a plurality of datasets. To give a concrete example, WEF (2015) reports that IBM has recently developed a comprehensive analytics tool called the Reuse Selection Tool, whose range of data input includes “product engineering and material data, modularity and reuse potential data, regulations and financial data such as market price and cost of remanufacturing, but also crucially the supply and demand data at various product taxonomy levels” (p. 26). Fundamental to computational modeling and the correlation of heterogeneous forms and sources of intelligence are parameters. They establish connections between various factors or variables, while adding numerical weightings to these connections. With their help, abundant, heterogeneous, and incommensurable datasets can become integrated into a single model that can recommend a single decision for action. As AI technology produces increasingly complex and dynamic models, parameters form “evolutionary variables that enter and exit relations with other parameters” (Parisi, 2013: 105) and they provide the programming capacities to “calculate potential conditions of relationality and change” (p. 106).

From a deontological perspective, responsibility plays out in the design and configuration of algorithms and their parameters (Burke, 2019). Parameters are articulated to indicators, which conceptually define and indicate “sustainability.” For instance, they measure and account for ecological impacts, such as energy and water expenditures, recovery and recycling of metals and plastics or rare earths that are kept “in the loop” via the circular economy. Models weigh such impacts against one another as well as against economic goals. In this regard, a deontological critique might highlight that models often encode sustainability within a cost-benefit analysis framework rather than as an issue of finite planetary capacities and resources per se. Hence, whereas optimization (Halpern, 2020) in recycling suggests a commitment to making waste mining more efficient and effective in terms of material recovery—closing the loop of the circular economy, so to speak—the term ambiguously also connotes reconciling objectives related to sustainability and profit. The priorities of this balance are expressed in parametric design. Optimization effectuating limited incremental reform rather than systemic transformation (Green, 2019) may be considered greenwashing by algorithmic means—a betrayal of the obligation to truly take responsibility for planetary presents and futures.

However, the issue of responsibility appears more complicated in the light of the fact that translating the value of “sustainability” into parameters evokes certain questions about the limitations of quantification and automation. An indicator such as “recovery rate” in a reverse logistics model may be hard to determine and forecast due to the complexity of underlying parameters, including “product-dependent” parameters (e.g. the value of disposed products, their capability to be repurposed or sold in secondary markets) as well as “social” parameters (e.g. habits, awareness about recycling, and consumer attitudes) (Tsouflas et al., 2009). The epistemic uncertainty intrinsic to such parameters is negated in the case of fuzzy logic applications that account for empirical liminal thresholds by producing calculations that factor in degrees of doubt and lack of information (Paul et al., 2021; Zhang et al., 2021). Furthermore, normative indeterminacy, constituting another facet of uncertainty, inhabits the ethical dilemmas of Multi-Criteria Decision-Making approaches. For instance, there is no single way to compare the respective ecological costs of “raw materials, energy and water consumption, solid and liquid wastes, emissions of noise and pollutants to the air, water and soil, and health and safety issues” (Bufardi, 2009: 76). Again, along with intuition and subjective judgment, fuzzy logic is used to translate imprecise, qualitative judments into quantitative values and weigh incommensurable indicators and objectives (Amoore, 2020; Efendigil et al., 2008: 273). For Derrida (2002), the undecidable resides not simply in “the oscillation or the tension between two decisions” but in the experience of having to bring within the “order of the calculable and rule” that which is “foreign and heterogeneous” to it. To decide in the face of the undecidable may be considered an act of taking responsibility. Yet, the process of automated, algorithmic decision-making in the smart circular economy threatens to undo this. For one, algorithmic decision-making can imply the “de-responsibilisation of human actors, or a tendency to ‘hide behind the computer’” whose outcomes are placed beyond doubt and seen as correct by default (Mittelstadt et al., 2016). In the context of the smart circular economy, this would imply the depoliticization of the value of sustainability, along with the demise of responsibility borne by human actors, as Derrida considers it. The issue appears worse in the light of the coalescence between the opacity surrounding algorithms’ code and operation, on the one hand, and institutional opacity, on the other (Burrell, 2016). The latter shrouds environmental standards that evoke the value of sustainability but refrain from associating it with strict, publicly-known norms (Easterling, 2014).

Furthermore, as Gordon et al. (2022) write, the process of associating lines of code with values constitutes a more extensive, uncertain process of translation. A value such as “sustainability” is performed through a host of “material actions and interactions conducted in its name” (p. 4). Yet such translation is uncertain and indeterminate, produced in and through the entire set of material conditions, practices, and effects belonging to algorithmic operation, such as the selection and labeling of data, experiments with parameters, self-learning behaviors, and effects of algorithms “in the wild.” In the light of the contingent operation of the assemblage in excess of the values attributed to code, algorithms do not just negate uncertainty but also double it in the sense that they generate new, contingent realities, practices, and struggles (Bucher, 2018; Esposito, 2022; Parisi, 2013).

The uncertainty of translation arguably problematizes and displaces deontological takes on responsibility. Bennett (2010) contends that the fact that agency unfolds through human-nonhuman assemblages inhibits assigning responsibility to any single actor. She takes issue with “blame politics” because the entanglements of persons and things imply that individuals, and by extent corporations, cannot bear full responsibility for their effects. Bennett's vision speaks as much to the liminality and agency of waste to become something else as to the uncertain translations of “sustainable” algorithms. Her view problematizes both corporate claims to responsibility and the expectation for corporate responsibility by others such as environmental activists or environmentally minded consumers and investors. Yet, as Bennett abandons the concept of responsibility, her position has been critiqued for undermining holding specific people or corporations accountable (and, paradoxically, for relying solely on a highly individualized type of personal ethics, despite the stated anti-anthropocentrism; Lemke, 2018; Washick et al., 2015). Given the contingency and uncertainty of the smart circular economy, the irresponsibility and opacity of the informal milieu could be replaced with novel forms of irresponsibility and opacity, associated with the algorithmically conducted smart circular economy. Therefore, rather than abandoning the concept of responsibility, I propose to rethink it in the context of smart infrastructures. I take cue from Amoore (2020), who designates a “cloud ethics” that “does not belong to an episteme of accountability, transparency, and legibility, but on the contrary begins with the opacity, partiality, and illegibility of all forms of giving an account, human and algorithmic” (p. 8). As one strategy for such an ethics, Amoore advises to consider not the source code as such but the absent, marginalized, and doubtful. She notes that algorithmic decision-making systems cultivate doubt within computation and yet “place the decision beyond doubt.” However, as her cloud ethics explores, would it be possible “to locate and amplify the doubtfulness dwelling within the partial fragments of the science itself?” Doubtfulness in her account signifies as much “hesitation, an uncertainty and a straying from calculable paths” as “a fullness and a plenitude of other possible incalculable paths” (p. 142).

In similar vein, the point of my comments regarding epistemic and normative uncertainty is not that we should relinquish computational modeling or any other method of “uncertain” calculation or assessment in order to be less speculative and more authoritative in our approach to waste. Instead, I argue that the design of the circular economy should allow for speculation beyond what current computational models afford. Expanding the speculative engagements with waste beyond the particular and exclusive gazes afforded by current models for “optimized” decision-making would mean having an expanded sense of matter's potential as well as accounting for the uncertainty and contingency, inherent in, and stemming from, algorithmic decision-making. For instance, how could circular economies account for the unknown or uncertain ecological harm that is associated with e-waste (or any of the materials deployed in production and recycling) entering wider current and future ecologies? Cubitt (2017: 119) notes that “no one knows” the full scale of ramifications that will be triggered by modern plasma screens, “which bombard phosphors with liberated ions for a hundred thousand hours before end of life,” once they enter the waste stream. Would a more ethical feedback loop be possible—one that opens up the possibility of taking responsibility not only for known risks but also unknown exclusions? What kind of human, machinic, or—more broadly conceived—more-than-human sensing and monitoring would this take (Gabrys, 2016a)? More so, could a mediated circular economy speculate about matter's potential to become while opening up rather than foreclosing possibilities for matter's becoming in the process of repurposing? If the current smart circular economy is designed to safeguard the formal sector's ontology of delimitation, what other circularities would be possible if decision-making algorithms would “follow the flow of matter,” like the metallurgist?

Conclusion

Drawing on critical data and algorithm studies, theories of waste, and empirical research, this paper has investigated ecological ethics for the circular economy in relation to datacentric and algorithmic technologies. I have discussed the ways in which the rationalities, affordances, and dispositions of such technologies perform and displace notions of corporate environmental responsibility and transparency. At stake is a slippage between response-ability, namely the ability to respond and act, on the one hand, and corporate responsibility, on the other. That is, smart technologies such as AI, Internet of Things, and Big Data promise corporate responsibility by facilitating response-ability. Yet, when considered critically, they enact surveillance and control while preserving an order, rooted in an ontology of delimitation, that reflects corporate interests and a restrictive distribution of agency. Following this approach, the corporation is the privileged decision-making agent in recycling processes and the authority over, if not owner of, users’ broken devices and waste. Corporate responsibility is associated with transparency toward the public, yet the transparency of corporate environmental sustainability reports exists alongside the black-boxing of not just electronic hardware, which users have no right to repair, but also the algorithms that guide “sustainable” corporate praxis. By comparison, the informal sector (in some respects) distributes the ability to respond, and material agency is probed and “followed” in processes of refurbishment, tinkering, and piracy, rather than being controlled. Hence, the informal sector undermines datafication, control, and (corporate) transparency but permits situated, tactical, and tacit response-ability by heterogeneous actors, however risky the involvement is. Moreover, while the informal sector is denunciated for its obvious lapses of responsibility, the smart circular economy displaces responsibility. This is so because algorithmic performances of sustainability are rife with uncertainties stemming from the translation of such a value into a host of material conditions, practices, and effects, from data labeling to parametric experimentation. Taking responsibility for waste and for planetary presents and futures is not achieved by the current investment in algorithmic control and smart circularity, which do not simply annihilate problems of waste, pollution, energy, and resource finiteness. Instead, I argue that the design of the circular economy should allow for speculation beyond what current computational models afford. Expanding the speculative engagements with waste beyond the exclusive gazes afforded by current “optimized” models would mean having an expanded sense of matter's potential as well as accounting for the uncertainty and contingency, inherent in, and stemming from, algorithmic decision-making.

My argument deviates from the perhaps more common critique of corporations not taking up enough responsibility by problematizing the ways in which corporate responsibility endows corporations with agency. In the context of the smart circular economy, corporate agency without oversight is reflected in algorithmic configuration such as the choice of parameters in decision-making models. What is obfuscated is the doubtfulness that inhabits the translation of the value of sustainability into algorithmic operations that encompass a host of material practices and effects. Having analyzed how the smart circular economy performs and displaces “responsibility” and “transparency,” I am left with the questions: without simply shifting responsibility to consumers or public entities, can my critique enable another position than calling for heightened corporate responsibility? And what could a more speculative form of transparency look like? Rather than demanding more “certain” knowledge about waste and pretending that smart systems will give us just that, I suggest that taking responsibility for present and future planetary conditions might involve different, more speculative systems and cultures that care for waste.

Footnotes

Acknowledgments

I would like to thank research assistant Kelvin Wu for his excellent help with the fieldwork underpinning this study.

Declaration of conflicting interests

The author(s) declared no potential conflicts of interest with respect to the research,authorship,and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research,authorship,and/or publication of this article: This work was supported by the Chinese University of Hong Kong,(grant number Direct Grant 4052253).

ORCID iD

Rolien Hoyng

References

Agostinho

(2019) The optical unconscious of big data: Datafication of vision and care for unknown futures. Big Data & Society 6(1): 1–10.

Amoore

(2020) Cloud Ethics: Algorithms and the Attributes of Ourselves and Other. Durham and London: Duke University Press.

Apple (2018) Environmental Responsibility Report. Retrieved from: https://www.apple.com/hk/en/environment/

Banasik

Bloemhof-Ruwaard

Kanellopoulos

, et al. (2018) Multi-criteria decision making approaches for green supply chains: A review. Flexible Services and Manufacturing Journal 30: 366–396.

Bell

(2019) Place, people and processes in waste theory: A global south critique. Cultural Studies 33(1): 98–121.

Bennett

(2010) Vibrant Matter: A Political Ecology of Things. Durham: Duke University Press.

Bisschop

(2012) Is it all going to waste? Illegal transports of e-waste in a European trade hub. Crime, Law and Social Change 58(3): 221–249.

Bucher

(2018) If/Then: Algorithmic Power and Politics. New York: Oxford University Press.

Bufardi

(2009) Environmental criteria in a MCDM context. In: Wang

H-F

(ed) Web-based Green Products Life Cycle Management Systems: Reverse Supply Chain Utilization. Hershey: IGI Global, pp.74–92.

10.

Burke

(2019) Occluded Algorithms. Big Data and Society 6(2). Available at: https://journals.sagepub.com/doi/10.1177/2053951719858743

11.

Burrell

(2016) How the machine ‘thinks’: Understanding opacity in machine learning algorithms. Big Data & Society 3(1): 1–12.

12.

Cubitt

(2017) Finite Media: Environmental Implications of Digital Technologies. Durham: Duke University Press.

13.

Dauvergne

(2020) AI in the Wild: Sustainability in the Age of Artificial Intelligence. Cambridge, MA: MIT Press.

14.

Davis

J-M

Akese

Garb

(2019) Beyond the pollution haven hypothesis: Where and why do e-waste hubs emerge and what does this mean for policies and interventions? Geoforum; Journal of Physical, Human, and Regional Geosciences 98: 36–45.

15.

Deleuze

Guattari

(2005) A thousand Plateaus: Capitalism and Schizophrenia. Minneapolis: University of Minnesota Press.

16.

Derrida

(2002) Acts of Religion. London: Routledge.

17.

Douglas

(1984) Purity and Danger: An Analysis of the Concepts of Pollution and Taboo. London and New York: Routledge.

18.

Easterling

(2014) Extrastatecraft: The Power of Infrastructure Space. London: Verso.

19.

Efendigil

Önüt

Kongar

(2008) A holistic approach for selecting a third-party reverse logistics provider in the presence of vagueness. Computers & Industrial Engineering 54(2): 269–287.

20.

Ellen Macarthur Foundation (EMAF) (2018) Circular Consumer Electronics: An Initial Exploration. Retrieved from: https://www.ellenmacarthurfoundation.org/publications/circular-consumer-electronics-an-initial-exploration.

21.

Ellen MacArthur Foundation (EMAF) (2019)) Artificial Intelligence and the Circular Economy - AI as a Tool to Accelerate the Transition. Retrieved from: https://www.ellenmacarthurfoundation.org/publications/artificial-intelligence-and-the-circular-economy.

22.

Esposito

(2022) Artificial Communication: How Algorithms Produce Social Intelligence. Massachusetts: MIT Press.

23.

European Environment Agency (EEA) (2017) Circular by Design: Products in the Circular Economy. Publications Office of the European Union. ISBN 978-92-9213-857-8 doi:10.2800/860754.

24.

Fazi

(2019) Digital aesthetics: The discrete and the continuous. Theory, Culture & Society 36(1): 3–26.

25.

Forti

Baldé

Kuehr

, et al. (2020) The Global E-waste Monitor 2020: Quantities, Flows and the Circular Economy Potential. United Nations University (UNU)/United Nations Institute for Training and Research (UNITAR) – co-hosted SCYCLE Programme, International Telecommunication Union (ITU) & International Solid Waste Association (ISWA), Bonn/Geneva/Rotterdam.

26.

Friant

Vermeulen

Salomone

(2020) A typology of circular economy discourses: Navigating the diverse visions of a contested paradigm. Resources, Conservation and Recycling 161: 104917.

27.

Gabrys

(2013) Digital Rubbish: A Natural History of Electronics. Ann Arbor, MI: University of Michigan Press.

28.

Gabrys

(2016a) Program Earth: Environmental Sensing Technology and the Making of a Computational Planet. Minneapolis: University of Minnesota Press.

29.

Gabrys

(2016b) Re-thingifying the Internet of Things. In: Starosielski

Walker

(eds) Sustainable Media: Critical Approaches to Media and Environment. New York and London: Routledge, pp.180–195.

30.

Galloway

(2004) Protocol: How Control Exists after Decentralization. Cambridge MA and London: MIT Press.

31.

Gille

(2010) Actor networks, modes of production, and waste regimes: Reassembling the macrosocial. Environment and Planning A: Economy and Space 42(5): 1049–1064.

32.

Goodall

Richard

Andrew

(2019) A data-driven simulation to support remanufacturing operations. Computers in Industry 105(2019): 48–60.

33.

Gordon

Rieder

Sileno

(2022) On Mapping Values in AI Governance. Computer Law & Security Review 46. https://doi.org/10.1016/j.clsr.2022.105712.

34.

Green

. “‘Good’ isn’t Good Enough. 2019, access September 5, 2022. https://www.benzevgreen.com/wp-content/uploads/2019/11/19-ai4sg.pdf

35.

Halpern

(2020) Planetary intelligence. In: Roberge

Castelle

(eds) The Cultural Life of Machine Learning: An Incursion into Critical AI Studies. Cham: Palgrave Macmillan, pp.227–256.

36.

Hogan

(2018) Big data ecologies. Ephemera: Theory and Politics in Organization 18(3): 631–657.

37.

Hoyng

(2018) Logistics of the accident: E-waste management in Hong Kong. In: Neilson

Rossiter

Samaddar

(eds) Logistical Asia: The Labor of Making a World Region. Singapore: Palgrave McMillan, pp.199–220.

38.

Hoyng

(2019) Aggregations of the opaque: Rethinking datafication and e-waste.” First Monday 4, no: 4. https://firstmonday.org/ojs/index.php/fm/article/view/9866/7747

39.

Hoyng

(2022) Toward a viral circular economy? Affections of data and waste. In: Hawkins

Rossiter

(eds) Contagion Design: Labor, Economy, Habits, Data. London: Open Humanities Press, pp.162–174.

40.

Huisman

Botezatu

Herreras-Martinez

, et al. (2015) Countering WEEE illegal trade (CWIT) summary report, market assessment, legal analysis, crime analysis and recommendations roadmap. Countering WEEE Illegal Trade (CWIT) Consortium. Retrieved from: https://collections.unu.edu/view/UNU:6359.

41.

Isenhour

Reno

(2019) On materiality and meaning: Ethnographic engagements with reuse, repair & care. Worldwide Waste: Journal of Interdisciplinary Studies 2(1): 1–8.

42.

Ivakhiv

(2014) Beatnik brothers? Between Graham harman and the deleuzo-whiteheadian axis. Parrhesia 19: 65–78.

43.

Jackson

(2014) Rethinking repair. In: Gillespie

Boczkowski

(eds) Media Technologies: Essays on Communication, Materiality and Society. Cambridge MA: MIT Press, pp.221–239.

44.

Jackson

Krieshok

Pompe

(2012) Repair Worlds: Maintenance, Repair, and ICT for Development in Rural Namibia. Computer Supported Cooperative 2012, February 11–15, 2012, Seattle, Washington, USA. Retrieved from: https://dl.acm.org/doi/10.1145/2145204.2145224.

45.

Laser

Stowell

(2020) Thinking like Apple’s recycling robots: Toward the activation of responsibility in a postenvironmentalist world. Ephemera Journal 20(4): 163–194.

46.

Lemke

(2018) An alternative model of politics? Prospects and problems of Jane Bennett’s vital materialism. Theory, Culture & Society 35(6): 31–54.

47.

Lury

(2009) Brand as assemblage. Journal of Cultural Economy 2(1–2): 67–82.

48.

Massumi

(2015) Ontopower: War, Powers, and the State of Perception. Durham: Duke University Press.

49.

Mittelstadt

Allo

Taddeo

, et al. (2016) The ethics of algorithms: Mapping the debate. Big Data and Society 3(2): 1–21.

50.

Nobre

Tavares

(2020a) Assessing the role of big data and the internet of things on the transition to circular economy: Part 1. Johnson Matthey Technol. Rev 64(1): 32–41.

51.

Nobre

Tavares

(2020b) Assessing the role of big data and the internet of things on the transition to circular economy: Part 2. Johnson Matthey Technol. Rev 64(1): 19–31.

52.

Offenhuber

(2017) Waste is Information: Infrastructure Legibility and Governance. Cambridge, MA: MIT Press.

53.

Pagoropoulosa

Pigossoa

McAloonea

(2017) The emergent role of digital technologies in the Circular Economy: A review. The 9th CIRP IPSS Conference: Circular Perspectives on Product/Service-Systems. Retrieved from: https://www.sciencedirect.com/science/article/pii/S2212827117301452.

54.

Parikka

(2015) A Geology of Media. Minneapolis: University of Minnesota Press.

55.

Parisi

(2013) Contagious Architecture: Computation, Aesthetics, and Space. Massachusetts: MIT Press.

56.

Paul

Shukla

Paul

, et al. (2021) Sustainable supply chain management and multi-criteria decision-making methods: A systematic review. Sustainability 13: 7104–7132.

57.

Posner

(2021) Supply chain. In: Nanna Bonde Thylstrup

Ring

Veel

D’Ignazio

(eds) Uncertain Archives: Critical Keywords for Big Data. Cambridge, MA: MIT Press, pp.493–499.

58.

Puig de la Bellacasa

(2017) Matters of care: Speculative ethics in more than human worlds. Minneapolis: University of Minnesota Press.

59.

Rossiter

(2016) Software, infrastructure, labor: A media theory of logistical nightmares. New York: Routledge.

60.

Ruppert

(2012) Category. In: Lury

Wakeford

(eds) Inventive Methods: The Happening of the Social. London: Routledge, pp.36–47.

61.

Terranova

(2004) Network Culture: Politics for the Information Age. London: Pluto Press.

62.

Thylstrup

(2019) Data out of place: Toxic traces and the politics of recycling. Big Data & Society 6: 1–9.

63.

Tsing

(2009) Supply chains and the human condition. Rethinking Marxism 21(2): 148–176.

64.

Tsing

(2015) The Mushroom at the End of the World: On the Possibility of Life in Capitalist Ruins. Princeton, N.J.: Princeton University Press.

65.

Tsoulfas

Pappis

Karacapilidis

(2009) Applying web-based collaborative decision-making in reverse logistics: The case of mobile phones. In: Wang

H-F

(ed) Web-based Green Products Life Cycle Management Systems: Reverse Supply Chain Utilization. Hershey: IGI Global, pp.401–415.

66.

Turner

Tam

(2022) Moving from risky to response-able care. Antipode 54(3): 914–933.

67.

United Nations environmental Program Finance Initiative (UNEP FI) (2020) Financing Circularity: Demystifying Finance for Circular Economies .

68.

Washick

Wingrove

Ferguson

, et al. (2015) Politics that matter: Thinking about power and justice with the new materialists. Contemporary Political Theory 14(1): 63–89.

69.

World Economic Forum (WEF) (2015) Intelligent Assets Unlocking the Circular Economy Potential. REF 081215 https://www.weforum.org/reports/intelligent-assets-unlocking-the-circular-economy-potential.

70.

World Economic Forum (WEF) (2018) Recovery of Key Metals in the Electronics Industry in the People’s Republic of China an Opportunity in Circularity .

71.

Xavier

Ottoni

Lepawsky

(2021) Circular economy and e-waste management in the americas: Brazilian and Canadian frameworks. Journal of Cleaner Production 297: 126570.

72.

Zhang

Venkatesh

Liu

, et al. (2019) Barriers to smart waste management for a circular economy in China. Journal of Cleaner Production 240, Article 118198.

73.

Zhang

Wang

, et al. (2021) An integrated multicriteria decision-making approach for collection modes selection in remanufacturing reverse logistics. Processes 9: 631–647.