This post was written by Veronica Fanzio
One query on ChatGPT needs almost 10 times as much electricity as a Google […] search, according to a study by Goldman Sachs’ […]. In April, Ami Badani, chief marketing officer of British chip designer Arm […], said data centers powering AI chatbots such as OpenAI’s ChatGPT account for 2% of global electricity consumption. That demand, Badani said, could eventually slow down AI progress.
In July, Google said its carbon emissions have risen by 48% since 2019, mostly due to energy consumption by data centers and supply chain emissions. The company’s carbon emissions were up 13% year over year in 2023, according to its 2024 Environmental Report.
source: Qz.com
In a world literally set aflame, and where development roadmaps seem to be governed by the Sustainable Development Goals set by the UN for 2030, it is clear that Big Tech companies need to take immediate action to lower their carbon emissions, and that AI is now at the centre of the conversation. As available data shows, AI is increasingly causing environmental damage: its large infrastructure is founded on the extraction of raw materials, electricity, and myriad polluting elements embedded in its complex supply chain.
A rising wave of AI ethics is focusing on this precarious relationship between AI and sustainability, stressing how the ubiquitousness of this technology makes its environmental impact unignorable (van Wynsberghe, 2021). AI operates at the advantage of an era characterised by economic hyper-growth (ALLIOUI & Mourdi, 2023) where businesses playing in the global economic landscape are capable of (and are incentivised to) reach a +40% annual growth rate (weforum.org).
In this economic environment, a growing public sentiment is turning toward degrowth (Schmelzer, Vetter, & Vansintjan, 2022). However, mainstream media, backed by Big Tech’s marketing efforts, paint a picture of steadfast techno-solutionism and the cause of the problem is offered as a solution; do a quick search on the internet for the query ‘Sustainable AI’, and corporate ads (often operating in the Software-as-a-Service industry) and consultancy blogs flood the screen with optimistic proposals and statements.
This consumerist discourse is imbued with positivist rhetoric that paints AI and digital technologies as our exclusive saviour but, behind the scenes, there are uglier truths. Tech giants do not seem to want to backtrack, nor rethink their relationship with AI, as is clear from the newly signed agreement between Google, Amazon, Microsoft and Dominion Energy, Energy Northwest, Kairos Power, and other US-based nuclear energy contractors (qz.com). Within these agreements, for instance, Amazon blatantly defends its support of the “development of nuclear energy projects”, buying exclusive supply of ad-hoc built nuclear reactors, more precisely small modular reactors (SMRs).
While recognised as an effective replacement for fossil fuel-fired plants, these reactors require materials to be mined, supplied, and assembled. Additionally, experts are sceptical about their capability to fully supply energy for data centres, where operators would have to provide backup power (Lyman, 2024). It’s to be seen how that will support Big Tech’s goals of achieving carbon neutrality.
These agreements are in line with tech companies’ trend to externally promote sustainable growth – an oxymoron in the capitalist system – and the familiar tendency to greenwash their entire supply chain and end products. These are marketing techniques aimed at appealing to the conscious consumer, but they are also PR strategies: companies increasingly have to appease the demands of investors who are becoming sympathetic to environmental regulations and policies. In the meantime, they operate in an economic landscape increasingly informed by the financialisation of natural goods such as forests and water sources, as seen with the rise of Natural Asset Companies functioning to privatise the property of all natural processes in a (US-based) territory (Bersani, 2023).
Confronted with this reality, words from the Global Digital Cultures soirée of October 20, 2024, resonate with urgency: “We need an infrastructure of change,” claimed the University of Amsterdam’s Niels ten Oever, one that goes beyond the infrastructure of violence and profit-centred mindset that has directed technological innovations to this day (ibid.). Big Tech shaking hands with nuclear energy suppliers is a manual-perfect manoeuvre aligned with the trends identified by the Critical Infrastructure Lab: digital technology companies attempt to counter their environmental harm by trying to rely on more efficient energy, “over-engineering without tackling the actual issue” (Fieke Jansen). These propositions are enacted within a narrative framework that depicts AI as simultaneously an immaterial (therefore environmentally harmless) agent and the epicentre of future-proofing climate solutions, a discourse that mythologises AI and delays our collective margin of action (Jansen & Thorne, 2024).
Hegemonic epistemologies are difficult to contrast, and changemakers can be fairly daunted by such delays. However, dialoguing with engaged scholars and researchers can provide some comforting perspectives, urging us to inhabit the ruptures within traditional barriers between academic knowledge and popular engagement. For example, we can begin reorienting our efforts and rethink innovations in terms of “infrastructure of peace”, a concept adopted by Fernanda R. Rosa to articulate open, communal, “pluriversal” projects and enact a substantial paradigm shift from how technologies are developed and infrastructures are built.
Drawing from a decolonial critique of the concept of “Universal” proper of Western philosophical traditions (Grosfoguel, 2017), Fernanda R. Rosa invites us to embrace the Zapatistas’ political vision, where multiple worlds coexist and a hegemonic cosmology is just one of the many (Mignolo, 2018).
Rosa’s espousal of a pluriversal approach accounts for all beings, moving beyond not only Western ontologies but crossing the boundaries of the human experience to embrace non-human actors. In practice, this means accounting for communitarian standpoints and embedding them into a shared understanding of the object of study. Her fieldwork in the Tseltal and Zapoteco communities in Mexico (Rosa, 2022) is exemplary, as it understands local internet interconnection infrastructure through a bottom-up approach and studies the emergence of shared networks, which are built locally and invisible in the traditional internet routing system. Shared networks thus defy the internet technical standards that see networks as owned by autonomous system operators; accounting for these systems as internet networks broadens the conventional spectrum of internet governance actors (ibid.) opening up novel possibilities for design and policy-making surrounding internet infrastructures.
We could all become governance actors if we are capable of rethinking governance as a whole and enabling a multitude of voices in the conversation. It is imperative to remove the logic of monetization (among others) from the equation and counteract the worldviews imposed by a few powerful global actors. To begin with, we can welcome the transition toward a pluriversal, community-oriented internet where everyone, not corporations, decides the terms of our connectivity.
Author Bio
Veronica Fanzio holds a research master in Media Studies from the University of Amsterdam and likes to delve into all that surrounds communications, glancing at the intersection between digital technologies and the public sphere. Currently a digital marketer in the educational sector, they analyse the impact of digital technologies on everyday life, with a focus on digital transformation and governance. Their research and writings revolve around promoting responsible, ecological, and equitable technologies.
Works Cited
ALLIOUI, H., & Mourdi, Y. (2023). Unleashing the Potential of AI: Investigating Cutting-Edge Technologies That Are Transforming Businesses. International Journal of Computer Engineering and Data Science (IJCEDS), 3(2), 1–12. Retrieved from https://ijceds.com/ijceds/article/view/59
Bersani, M. (2023). La rivoluzione della cura. Uscire dal capitalismo per avere un futuro. Edizioni Alegre.
Grosfoguel, R. (2017 [2006]). Decolonizing Western universalisms: Decolonial pluri-versalism from Aime Cesaire to the Zapatistas. In J. M. Paraskeva (Ed.), Towards a just curriculum theory: The epistemicide (1st ed., pp. 147–164). Routledge. https://doi.org/10.4324/9781315146904-6.
Jansen, F. & Thorne, M. (2024, October 15). Predatory Delay and Other Myths of “Sustainable AI”. AI Now Institute. https://ainowinstitute.org/publication/predatory-delay-and-other-myths-of-sustainable-ai
Lyman, E. (2024, April 30). Five Things the “Nuclear Bros” Don’t Want You to Know About Small Modular Reactors. The Equation. https://blog.ucsusa.org/edwin-lyman/five-things-the-nuclear-bros-dont-want-you-to-know-about-small-modular-reactors/
Mignolo, W. (2018). Foreword. On Pluriversality and Multipolarity. In Reiter, B. (Ed.), Constructing the Pluriverse: The Geopolitics of Knowledge. Duke University Press
Rosa, F. R. (2022). From Community Networks to Shared Networks: the Paths of Latin-Centric Indigenous Networks to a Pluriversal Internet. Information, Communication & Society, 26(11), 2326–2344. https://doi.org/10.1080/1369118X.2022.2085614
Schmelzer, M., Vetter, A. & Vansintjan, A. (2022). The Future is Degrowth: A Guide to a World Beyond Capitalism. Verso Books.
van Wynsberghe, A. (2021). Sustainable AI: AI for sustainability and the sustainability of AI. AI Ethics 1, 213–218. https://doi.org/10.1007/s43681-021-00043-6
