Send message to Author

Short abstract

This paper frames mass-produced low-quality AI-generated content, known as AI slop, as digital excess, contributing to critical debates on AI’s environmental costs and highlighting how platform capitalism amplifies wasteful forms of digitalisation that intensify ecological harm.

Long abstract

The low-quality, repetitive, and, at times, dubious AI-generated text, images, videos, and music, known as AI slop, is increasingly spreading on online platforms. On YouTube, nearly one in ten of the fastest-growing channels publishes exclusively AI-generated content (Bharadia, 2025). In publishing, Amazon’s response to mass AI-generated submissions has been setting a limit of three self-published Kindle books per author per day (Brodsky, 2024). Spotify removed 75 million spam tracks in 2025 after a surge of AI-generated music (Milmo, 2025), while on platforms such as Pinterest, the flood of AI-generated imagery has led to users questioning whether it remains usable (Rowe, 2025).

This paper conceptualises AI slop as a form of digital excess, amounting to something more than necessary, wasteful, and superfluous (Olsson et al., 2023; Vigren et al., 2026). While debates on AI’s environmental footprint often focus on energy-intensive model training and data centers, this paper shifts attention to the mass generation of AI content, driving greater storage, processing, and distribution demands and increasing energy and water consumption. Crucially, this excess is deeply embedded in the political economy of platform ecosystems. Rather than tackling content overload, major platforms encourage rapid, high-volume production, with algorithmic visibility and monetisation models that reward scale rather than quality, effectively institutionalising digital waste. By framing AI slop as digital excess, the paper contributes to critical debates on AI’s environmental costs and highlights how platform capitalism amplifies wasteful forms of digitalisation that intensify ecological harm.

AI slop; digital excess; generative AI; environmental costs; platform capitalism

Send message to Author

Short abstract

Based on empirical research with families using YouTube Kids, this paper explores digital noise as informational waste. Repetitive, low-value videos generated by recommendation systems create a noisy media environment that families must constantly navigate and filter in everyday viewing practices.

Long abstract

This paper examines digital noise as a form of informational waste through ethnographic observations of everyday engagements with YouTube Kids. Although the platform is designed as a curated and safe environment for children, parents frequently describe encountering streams of repetitive, low-value videos that feel overwhelming and difficult to filter. During interviews and observation sessions conducted with families using YouTube Kids in domestic settings, participants repeatedly referred to this content as “noise”: videos that constantly demand attention yet provide little narrative or educational substance. The study combines ethnographic interviews with parents, observation of children’s viewing practices, and discussions with small-scale content creators who produce videos for the platform. These encounters reveal how families actively navigate algorithmically generated streams like skipping, blocking, or redirecting content, while creators describe experimenting with small variations of similar videos in response to shifting recommendation metrics.

Rather than treating digital waste as a downstream effect of discarded devices or obsolete infrastructures, the paper shows how informational waste emerges upstream through algorithmic systems that incentivise continuous production and micro-variation. The resulting accumulation of repetitive videos produces a noisy viewing environment that families must constantly manage.

Keywords

digital noise, informational waste, algorithmic recommendation, platform economies, attention economies

Send message to Author

Short abstract

Drawing on the case automated river navigation, I reconceptualise waste not as a material residual but as a necessary temporal-discursive construct for automation. Reading automation through the lens of waste elucidates why so many speculative technologies create the problems they profess to solve.

Long abstract

This paper draws on the case of Rhine river navigation to examine how efficiency discourses are bound up in particular spatio-temporal constructs of ‘waste’. While 19th century river engineers saw as their economic and moral mission to tame the unruly river through rectification projects that rationalized space, the contemporary development of remote navigation is similarly couched in twin discourses of productivity and environmental sustainability but seeks instead a rationalization of time. My paper departs from dominant notions of waste as material residual downstream from production to instead conceptualise waste as a temporal construct that forms the discursive starting point for automation technology. The very idea of remote navigation, I argue, arises from – and only makes sense through - a particular construction of the shipping labour process as wasteful. Remote navigation envisions screen-mediation as an abstraction technology to make skippers’ labour power spatio-temporally interchangeable and distributable across fleets, allowing a recuperation of ‘wasted’ labour power through digital re-embodiment. Automation in this sense seeks not to eliminate labour but to recuperate labour time perceived as wasted. Drawing on three months of ethnographic field work at a remote navigation center, I then show how the attempt to abstract a collective labour process creates new forms of compensatory labour as remote skippers negotiate the reconfigured sociality required from the spatial severance of the ship as a social body. Ultimately, I argue that reading automation through the lens of waste helps elucidate why so many speculative technologies create the problems they profess to solve.

Send message to Author

Short abstract

Automation in Hamburg’s port relies on simulation and data systems that frame humans as “sources of disturbance.” This paper argues automation produces “wasted” human capacity by rendering some skills superfluous while requiring new ones, generating compensatory labour to stabilise its limits.

Long abstract

Each year, around eight million containers pass through the Port of Hamburg, a key node in global supply chain capitalism. Their circulation is coordinated through sociotechnical infrastructures that are still in the process of undergoing digital transformations that promise efficiency and seamlessness. Central to these developments are simulation models and data-driven training environments designed to anticipate and optimise logistical futures. This paper draws on the case of container terminal automation in Hamburg to examine how discourses of efficiency are bound up with particular constructions of the human body as something that “wastes” time and space due to its discursively produced “inferiority” compared to technology. The very project of automation, I argue, arises from, and is rendered intelligible through, a particular construction of human labour as inefficient and excessive. In industry discourse, the human appears as a “source of disturbance”, positioned as friction within data-driven systems. Certain skills are rendered superfluous while others are newly required, producing “wasted” human capacity alongside new dependencies. Drawing on empirical research at the Container Terminal Altenwerder, and through the perspective of “waste”, I show how digitalisation-driven automation reconfigures the labouring body and generates forms of compensatory labour required to stabilise its own limits.

Send message to Author

Short abstract

This paper examines AI data center complexes as new ecological and aesthetic wastelands. By attending to the resilient, often invasive plant species inhabiting these environments, I argue for data center landscapes as expanding wastelands harboring relentlessly vital life on a transformed Earth.

Long abstract

This paper chronicles encounters with ruderal plant species (weeds) growing in and around data center campuses. The analysis draws on field observation and site documentation conducted in northern Virginia’s “Data Center Alley” and at massive new construction projects on former mining sites in rural Appalachia. I situate these resilient plants’ lives within the wider context of data centers’ effects on soil contamination and groundwater levels and within the region’s history of ongoing extraction. I argue that the data center campus, as a site of intensive energy consumption and waste production, presents an exemplary late Anthropocene wasteland. The data center campus and its spontaneous, often invasive vegetation emerge here as successors to earlier industrial wasteland landscapes, provoking similar aesthetic negativity or indifference. In attending to the lives finding their way in these expanding wastelands, this paper asks how we might begin to reconcile ourselves to the new landscapes of a transformed Earth. It further asks whether an unchecked data center construction boom and a possible AI bubble might instead leave us with future weedy landscapes surrounding abandoned, half-built, or repurposed data centers. The presentation will include a visual element in the form of a weed herbarium containing specimens collected from data center research sites.

Send message to Author

Short abstract

This proposal forwards a novel understanding of 'data waste' through the lens of political ecology. This is achieved through a literature review of the convergent definitions of 'data waste', while leveraging Ireland as a 'waste frontier', and proposing a new field of 'Critical Data Waste Studies'.

Long abstract

In academic literature, the concept of data/digital waste has a wide-ranging understanding dependent on the fields in which it is leveraged. Convergent terms such as 'digital pollution', 'data emissions', 'data exhaust', 'data spills', 'AI Fumes', 'digital smog', etc. (Hasselbalch, 2022, Hogan, 2024), represent the emergence of a necessary evolution of the field of 'Critical Data Studies' (Edwards et al., 2024), to capture the environmental dimension of data materiality, that has yet to coherently or formally coalesce into a field in of itself. While efforts have been made to unify and coherently marry these convergent conceptualisations (Lucivero et al., 2020; Hasselbalch, 2022), these are often wedded to the limiting frames of 'sustainability' and policy solutionism, which often neglect a comprehensive interrogation of the unequal systems and exchanges, such as Ecological Imperialism (Pedregal and Lukić, 2024), that underpin the environmental harms of the production, storage and management of data waste writ large. Therefore, this proposal offers three novel contributions: One, a literature review of the convergent definitions of 'data waste' through the theoretical lens of political ecology. Two, the exploration of Ireland as a 'waste frontier' (Liboiron and Lepawsky, 2022; Bresnihan and Brodie, 2023) to understand this unified configuration of 'digital waste', namely through the proliferation of data centres and associated infrastructures. And three, the proposal of a new field of 'Critical Data Waste Studies' to properly account for and interrogate the environmental harms of digital infrastructures and the political economy that deems this waste necessary for capital accumulation.

Send message to Author

Short abstract

This paper explores the application of Elder Albert Marshall's concept of 'two eyed seeing' as a strategy for navigating decolonial gestures in knowledge politics, with a focus on different aspects of 'waste' relating to water (as a resource) and different forms of knowledge.

Long abstract

Drought and data centre construction have refocused attention on the politics of water in ways that highlight the potential and complexity of decolonial knowledge practices such as ‘Two Eyed Seeing’ or Etuaptmumk (Reid et al, 2021). This knowledge heuristic highlights the importance of rendering indigenous and non-indigenous knowings and beings epistemically equivalent. Water politics and knowledge politics illustrate and challenge these approaches. In a paradigm shaped by colonial ideas of water as a resource, water becomes recast as scarce, valuable and a site of struggle - including around data centre construction . At the same time, other 'ways of seeing' position rivers as living, rights-bearing entities and floods as necessary and natural phenomena.

Not 'wasting' water provides a lever for opponents of data centres, but also evokes a longer colonial history. The prevention of waste casts water as a resource and parsimony as a core value, evident in flood prevention schemes that seek to capture water while also reducing risk. This singular view can also therefore 'waste' the embodied understanding of how to live with and among floods. At the same time, seeking to always include marginalised knowledge can fail to account for politics of silence and restraint that are also part of indigenous knowledge practice, challenging colonial conceptions of waste.

This talk explores conceptions of waste and knowledge through stories of two rivers in the dry plains of North America, the Red River of the North and the Old Man River.

Send message to Author

Short abstract

A collective walkshop exploring the panel's themes and emergent ideas in dialogue with the river Kistula.

Long abstract

This open panel session aims to facilitate conversations around the subject of digital waste and rivers, applying field research methods to develop situated perspectives towards river ecologies as political sites of environmental social justice and resistance against intensive extractive practices in digital infrastructure.

Considering humanitarian and environmental policies have failed to protect industrial mining practices and tech companies to colonise, extract and poison waterways in the 'Global South', and the intensive demand for water governance to cool expansive server centres, what alliances can be formed by learning from environmental activist movements, such as citizen science and the river rights movement, to challenge political representation of water as resource but instead to consider it within the framework of transcorporeality and non-human subjectivity.

The walkshop will utilise walking as methodology to mobilise post-human perspectives towards interconnected bodies (biological, ecological, geological and technological) and create a dynamic space for reflecting on various ontologies of waste, through engaging with rivers and the non-human world through sensory led activities, prompts and discussion.

To take part in this off-site activity - please email m.dovey@soton.ac.uk with your name for details.