Panel «Hybridising Environmentality» at STS-CH 2025

Convened by CML contributors Cristina Cochior, Guillemette Legrand, and Helen V. Pritchard at STS-CH Conference 2025 in Zurich, this panel inquires into the increasing reliance of environmentality on Big Tech infrastructure and the technical hybridisation of environmental computation with Artificial Intelligence (AI) in contexts such as Earth monitoring, modelling and policing. The promises and imaginaries of this integration of AI within existing environmental infrastructures raises questions about their technical but also economic, political, and material hybridisation. The development of high-resolution, universalist and fast-paced models toward the so-called green transition reconfigures the dynamics and influence of the private sector and Big Tech within environmentality and its imaginaries. In this panel, we are interested in investigating how environmentality is reframed through its increasing hybridisation with private interests, automation, AI imaginaries and anticipatory techno-solutionism. (Interests that often seek to obscure more situated or community based proposals). We invite papers that look at how existing practices that visualise, measure or organise our governance of environments––are being configured by this so-called “hybridity”.

This panel is interested in projects that document this ongoing hybridisation and also counter and propose other possible renders of new forms of environmentality, including community servers, permacomputing, and minor technologies.

In this panel, we would like to discuss topics related but not limited to the following:

The involvement and impact of Big Tech on environmental research and the green transition
Critical experimentation with hybrid models and AI/ML pipelines in environmental and climate research
Other forms of environmental computation to counter shoved-AI imaginary
Community practices and grassroots governances configuring digital infrastructures otherwise

With contributions by Cristina Cochior & Helen V. Pritchard, Guillemette Legrand & Joan Llort, Stella De Luca & Isabella Traeger, and Larissa Lenze.

Find all abstracts of the panel in the conference booklet.

Digital Twin Engine: synthetic imaginaries of Earth

Guillemette Legrand, Joan Llort

This paper discusses the shifting techno-politics of climate modelling practices through the increasing integration of AI processes into traditional (physical) models. This integration results in so-called hybrid models, or even in some cases, in supplanting climate models with deep learning weather and climate algorithms. We propose to look at this shift through the case of the Digital Twin Engine, an interface described as the orchestrator of the digital twins, configuring a new form of data handling and interactivity within the European project Destination Earth. We explore the computational modalities through which the Earth’s twin formulates and anticipates climate imaginaries through asking: What are the limitations of the Earth’s twinning? Are digital twins more clones or counterfeits? How does the Digital Twin Engine ‘twins’, what is excluded or misrepresented in this process, and how is this communicated back to the ones operating it? We will document the limitations of ‘twinning’ the Earth in the context of the Digital Twin Engine, through mapping its reliance on synthetic data to train AI algorithms, the mimicking of existing large climate models with their uncertainties, and finally, the gamification of Earth’s simulation through an engine interface. This paper documents but also speculates about the socio-technical imaginaries of the Digital Twin Engine in the anticipation of its full-scale launch.

Hybridity and fragmentation at the peripheries of digital infrastructures

Cristina Cochior, Helen V. Pritchard

This paper looks at how infrastructural scale shapes the epistemic valuations attached to “hybridity” and “fragmentation” in relation to the climate polycrisis. While large-scale Big Tech infrastructures present themselves as stable, coherent, and finished, small-scale grassroots digital infrastructures embrace situated and ad hoc practices.

This paper argues that this asymmetry does not stem from intrinsic qualities, but from divergent intentions to conceal or reveal infrastructural practices. We argue that large-scale infrastructures routinely incorporate bricolage, and improvisation, features typical of experimental practice, yet these are rendered invisible through discursive techniques that produce an illusion of unity and permanence. By contrast, the experimental and patchwork qualities of small-scale digital infrastructures, such as community run servers, are hypervisible. Fragmentation in small-scale infrastructure might be understood as a form of hybridity; not a failure to cohere, but a situatedness of technologies, knowledges, and temporalities. Yet because they don’t fit dominant infrastructural practices and aesthetics or models of scale, they’re often rendered marginal or “inefficient” and generally undervalued in climate policy and governance—by corporate and state actors.

How is infrastructural legitimacy performed and scaled, and how might fragmentation be reimagined as a site of possibility? Through disobedient policy analysis and practice based research our paper proposes a consideration of small-scale practices not as peripheral or incomplete versions of large-scale models, but as complex sites of attunement that challenge linear narratives of infrastructural progress and how we imagine, build, and justify infrastructural responses to ecological breakdown.