Literature Review & Rationale: 


The Environmental and Social Impact of Digital Technologies and Data Centres

The environmental and social impact of digital technologies, data centres, and digital devices is becoming an ever increasingly intersectional problem, with pledges for impact reduction and resilience being made at all scales: global aims (UN, 2024); national aims (UKGOV, 2024); and sectoral aims (UniversitiesUK, 2023). At an institutional level University of The Arts, London has committed to the significant reduction of its carbon emissions (measured in CO2 equivalent/CO2e, which is a simplified metric where any greenhouse gas (N2O, methane, hydrofluorocarbons, etc) are converted into equivalent CO2 so they can be easily compared) by 2040 (UAL, 2023a; UAL, 2023b). The university has committed to 92% reduction (ibid.) in scope 1+2 carbon emissions (scope 1+2 are emissions in direct control of the institution, such as heating and energy usage). With scope 3 (indirect, supply chain, etc) reductions coming a decade later in 2050 (ibid.). This ambitious target is in a way essential as universities integrate more digital systems, but it also highlights the environmental challenges tied to our expanding digital footprint.
These problems are vast — at an almost innumerable scale — and reduction pledges are made all the way through the human-wrought late-capitalist hierarchy (global, national, sectoral, institutional), but my feeling is that although large swathing actions are both justifiable and magnanimous, small measurable changes can happen at the individual personal level. These small actions of change have the capacity to leave us with the feeling of agency, and agency suggests hope (Zaki, 2024), rather than unsurmountable doom.
This literature review explores the wider global and situational conditions for this Action Research Project, and some of the contextual challenges, focusing on data centres, the hidden carbon costs of digital technologies, and the intersectional impacts of digital infrastructure. I conclude by acknowledging all of these intersecting, complex, hyper-object-like (Morton, 2016) problems, and state that the rationale for this project is to look at my individual contribution through my work at UAL and intend to do something about it.

Data centres (see Figure 1) are the hidden backbone of the modern digital landscape, storing everything from academic research to cloud storage of photos of cats. Yet, they come with a significant environmental cost. Recent studies estimate that global data centre energy consumption reached 196 terawatt hours (TWh) in 2020 (Mytton & Ashtine, 2022; Masanet et al., 2020). In addition, the networking systems that support these centres consumed 272 TWh in 2020, a 10% increase from 2015 (Malmodin et al., 2023). As demand for data transmission grows — up 600% from 2015 to 2022 (IEA, 2023) — power demand grows too, but interestingly there is not a direct relationship between gigabytes of data transferred and amount of energy consumed (Malmodin et al., 2024). This complex relationship needs further research and is out of scope for this project, and its on the cutting edge of network-based research. This already mentioned study by Jens Malmodin (2024) suggests there is not a linear causality between volume of data transferred and energy consumed, but it does suggest the storage of data uses energy, and that the cloud servers storing data have a permanent coefficient of energy usage.
Institutions like ours, UAL have turned to cloud computing to reduce their direct (scope 1+2) energy use. Cloud services are more energy-efficient due to their centralised nature, where large providers like our provider, Microsoft manage highly optimised data centres. However, as The National Archive (2023) notes, the carbon savings reported by these providers often rely on rough estimates, and there is little transparency about the actual emissions associated with these services. The embodied carbon of cloud infrastructure—considering everything from device manufacturing to waste disposal—remains significant, even if direct emissions are reduced (The National Archive, 2023).

Beyond the energy consumed by data centres and the transference of information, we must also consider the growing potential issue of ‘dark data’. Jackson and Hodgkinson (2024) describe this as the large volume of unused but stored data, which can be overlooked when calculating digital carbon footprints. This includes everything from old documents in cloud storage systems like OneDrive and Google Drive to archived media that are rarely, if ever, accessed. As Dourish (2017) argues, invisible digital storage has disconnected us from the immense physicality of information storage, fostering a kind of digital hoarding. In turn, this has escalated the demand for more and more storage capacity, with significant environmental consequences as described earlier. Furthermore, Siddik, Shehabi, and Marston (2021) highlight, U.S. data centres alone consumed 513 billion litres of water in 2018. Such figures bring attention to the hidden resource consumption of the digital world, which extends far beyond electricity use.
Anecdotally, UAL in 2022 doubled the limit of academic staff OneDrive storage from 500gb to 1000gb (1tb) — what impact could this have caused when doubling the 2740 (HESA, 2024) academic staff’s data limit? For context thats 1,370,000gb more of storage, or 1,370terrabytes or 1.37petabytes. This extra storage is unlikely to be actually used by all staff of course, but the availability is real.

The environmental impact of digital technologies is not felt equally to all people and environments. As Davis and Todd (2017), Davis (2023), and Nayar (2021) argue, the extractivist practices underpinning the manufacturing and production of digital devices disproportionately affect the global majority. The extraction of rare earth minerals needed for producing digital hardware often takes place in regions home to indigenous populations. This risks perpetuating colonialist environmental injustices, as the destruction of ecosystems and communities accompanies the growing demand for digital infrastructure (Davis, 2023).
Whyte (2016) discusses how this extractivist model echoes colonial practices, where the resources of marginalised communities are exploited for global technological advancement. This colonial dynamic is increasingly evident in the digital world, where the benefits of technological progress are often unevenly distributed. As Swift (2022) argues, mainstream environmental movements often fail to address these intersectional issues, leading to a form of climate activism that overlooks the voices and struggles of the most affected people and areas (MAPA), who are disproportionately affected by the impact of climate change (Greene-Dewasmes, 2023). Therefore impact that digital technologies (or anything for that matter) has, both impacts more, and goes overlooked in global majority populations.

Higher education institutions, such as UAL, can play a role in shaping the impact of digital technologies, ib both the dissemination of ideas and empirical actions. Barnett (2021) argues universities could become inclusive ecologies presenting the notion of the Ecological University, and questioning: How might higher education contribute to better futures for all in 2050? and How should higher education evolve to meet the challenges of this future? (ibid.). Higher education can of course contribute to a ‘better future’ by embracing human and nonhuman ecological sustainability and as part of that shaping and forging of responsible digital practices, but the unhelpful question remains — how? UAL, is of course attempting to do this, but as often is the way with anything intentionally radical but is actually institutionalised — it doesn’t go far enough.
Many institutions are focusing on ‘greening’ their IT systems, The National Archives (2023) internal study suggests that this alone is not enough. Instead, universities (and indeed all of us) must reduce hardware waste, use and improve the efficiency of the devices they use, how we do this, none of us are sure. Mike Berners-Lee (2020) suggests most of the CO2e impact is embodied within the manufacturing and the extracted raw materials of the device itself, not its use. This ‘embodied carbon’ of the digital devices themselves, and the environmental costs associated with their production, life-cycle, and disposal, can be overlooked in analysis.

While much research has focused on the energy consumption of data centres and networks there are still significant gaps. One key limitation is the lack of transparency about the empirical carbon emissions of cloud service providers. Companies like AWS (Amazon Web Services) and Microsoft Azure often provide only rough estimates of carbon savings, or none at all (Etsy, 2016). More reliable data on the carbon footprint of cloud services and the lifecycle of digital hardware is needed to form a clearer picture of their environmental impact.
Additionally, while much of the focus has been on energy use, the intersectional social and environmental consequences for the global majority remain under-explored.

Digital technologies, particularly data centres, appear to have a significant environmental and social impact that cannot be ignored. Universities like UAL are showcasing themselves as taking important steps toward reduction of impact, but as always more needs to be done to address the broader issues. As mentioned, the growing demand for digital infrastructure places a heavy strain on global resources, and the colonial dynamics of the manufacture of digital hardware exacerbate existing inequalities.

These issues go way beyond what is capable for a university, or indeed even a continent to solve, or reduce enough to preserve ongoing human existence. So in this project, we do not try. But I do want to know what is my rough, very rough, estimated impact through my university-based work using digital technologies. I teach predominantly digital software and physical hardware, using computers and devices, so I am unavoidably contributing to the specified impact of digital tools and their perpetual desire.
This project is in order to try and work out how I can personally reduce my impact and disseminate the above knowledge and encourage others to do so too.

.

---

References

Barnett, R. (2021) Towards the Ecological University: A Concept Note. [Online] Available at: https://ronaldbarnett.co.uk/Futures%20Project%20-%20concept%20note.pdf [Accessed 19 January 2025].

Davis, H. (2023) “Waiting in Petro-time,” Environmental Humanities, 15(3), pp. 52–64. Available at: https://doi.org/10.1215/22011919-10745979.

Davis, H. and Todd, Z. (2017) On the Importance of a Date, or Decolonizing the Anthropocene, Environmental Humanities, 16(4), pp. 761-780.

Dourish, P. (2017) ‘The Materialities of Information: Digital Hoarding and the Changing Spatialities of Data’. Technology and Culture, 58(4), pp. 3-5.

Etsy (2016) ‘Cloud Jewels: Estimating kWh in the Cloud’. [Online] Available at: https://www.etsy.com/codeascraft/cloud-jewels-estimating-kwh-in-the-cloud/ [Accessed 05 January 2025].

Greene-Dewasmes, G. (2023) Climate week messages from Most Affected People and Areas, World Economic Forum. Available at: https://www.weforum.org/stories/2023/10/mapa-voices-key-climate-week-messages-from-most-affected-people-and-areas/ (Accessed: January 20, 2025).

HEPI (2024) Decarbonising higher education – the investment challenge, HEPI. Available at: https://www.hepi.ac.uk/2024/06/21/decarbonising-higher-education-the-investment-challenge/ (Accessed: January 20, 2025).

Jackson, T. and Hodgkinson, I. (2024) ‘Decoding the Digital Carbon Footprint: Exposing the Global Data Challenge’, Chronicle.com. Available at: https://sponsored.chronicle.com/decoding-the-digital-carbon-footprint-exposing-the-global-data-challenge/index.html [Accessed 1 December 2024].

IEA (2023) ‘Global Data Centre Energy Consumption 2023’. International Energy Agency. [Online] Available at: https://www.iea.org/reports/global-data-centre-energy-consumption-2023 [Accessed 19 January 2025].

Masanet, E., Shehabi, A., & Koomey, J. (2020) ‘The Energy and Climate Impacts of Data Centres’. Nature Communications, 11(1), pp. 1-12.

Mytton, M. & Ashtine, A. (2022) ‘The Evolution and Energy Footprint of Data Centres’. Journal of Sustainable Technology, 15(2), pp. 124-137.

The National Archives (2023) ‘Digital Services and carbon emissions in the heritage sector: some preliminary findings – Archives sector’. Available at: https://www.nationalarchives.gov.uk/archives-sector/digital-services-and-carbon-emissions-in-the-heritage-sector-some-preliminary-findings/ (Accessed: 16 December 2024).

Nayar, J. (2021) Not So ‘Green’ Technology: The Complicated Legacy of Rare Earth Mining, Harvard International Review. Available at: https://hir.harvard.edu/not-so-green-technology-the-complicated-legacy-of-rare-earth-mining/ (Accessed: January 15, 2025).

Siddik, A., Shehabi, A., & Marston, S. (2021) ‘Water Consumption and Energy Use of Data Centres: The U.S. Context’. Journal of Environmental Impact Assessment, 37(3), pp. 245-260.

Swift, J. (2022) ‘Climate activism without intersectionality isn’t enough: Why we need intersectional environmentalism’, Black Women Radicals. Available at: https://www.blackwomenradicals.com/blog-feed/leahthomas [Accessed 19 January 2025].

UAL (2023a) Carbon Management Plan – realising a net-zero carbon institution by 2040. Available at: https://www.arts.ac.uk/__data/assets/pdf_file/0021/213852/UAL-CMP-v1272.pdf.

UAL (2023b) Environmental policy and management, UAL. Available at: https://www.arts.ac.uk/about-ual/climate-action-plan/change-the-way-we-operate/environmental-policy-and-management (Accessed: January 13, 2025).

Whyte, K.P. (2016) ‘Is it Colonial Déjà Vu? Indigenous Peoples and Climate Injustice’. In: Humanities for the Environment, pp. 102-119. Routledge.

Zaki, J. (2024) Hope for cynics: The surprising science of human goodness. Little, Brown Book Group.