Authors:
Melissa Gregg
How do technology users learn about the sustainability credentials of software and hardware products? Is there a role for designers to communicate the environmental factors involved in building and using a product to those buying it? In this column, I'll consider a handful of individuals and organizations trying to answer these questions, showing how everyday computing habits affect the finite carbon budget that allows us to remain on a livable planet. Before continuing, though, a point of clarification on sustainability, and the limits of the term green.
Many large technology companies have made public commitments to achieve net-zero targets in the next decade or so. Given the peculiarities of carbon emissions accounting, some even claim to have met this objective already, alongside business growth. To recap, net zero means that a business proposes to make a neutral impact on the environment, as measured by the global Greenhouse Gas Protocol. This formula provides disclosure guidelines to account for the many pollutants that are emitted from a company's operational infrastructure—from sourcing construction materials and powering facilities to the commuting footprint of employees. Calling out carbon dioxide and "equivalent" gases that contribute to atmospheric warming, the framework captures the noxious by-products that are the result of making and distributing physical products to consumers through various assembly and transportation channels.
For most of us, even if we care about the state of the planet, following the ins and outs of megawatts versus terawatts or the dangers hidden in supply chains are not preferred recreational pursuits. In this context, net zero can sound like a reassuring refrain, a sign that everything is going to be fine. It suggests that we can continue to do what we want (in this case, using computing devices), since tech companies will cancel out any side effects by buying carbon credits or investing in renewable energy.
Unfortunately, that approach isn't going to work. Even if all the renewable energy currently being set up goes toward computing, that pushes every other industry (not to mention our homes) on to carbon-intensive energy sources. More renewable energy exists today than there used to be, and that's great. But there are more energy consumers, too. And many of the products being released by hightech titans are driving up energy consumption (hello, generative AI). The problem with net-zero claims is the word net: If we aren't reducing demand for resources, we aren't addressing the reality that there aren't enough of them to go around.
Function Creep = Carbon Creep
Manufacturing new hardware is the largest carbon impact most technology products cause. A less obvious factor, and one that can increase this debt over time, is the constant push for software upgrades. The default for many computing devices today is to run upgrades in the background, so that we don't notice them. If we charge our phones overnight, the operating system often downloads and installs updates while we sleep. The features that are smuggled in with upgrades therefore don't often register as requiring greater processing requirements, because they are kept out of view (Meta adopted this approach when adding AI to consumer-facing apps this past year, to some confusion [1]).
Persistent function creep is one reason why average monthly household broadband consumption in the U.S. has grown from 9 GB in 2010 to 606 GB in 2024, according to the quarterly industry survey by OpenVault (https://openvault.com/). UX designer Tristin Oldani brought these numbers to my attention while we worked together on the Green Software team at Intel [2].
A lot has changed in just a few years when it comes to energy consumption on devices. When we started the Carbon Reduction and Green Software team in 2022, many of the companies producing popular software applications didn't have sustainability reporting structures in place. To get around this, Tristin took the initiative to create a scoring method to track the relative "greenness" of commonly used apps.
Cross-referencing publicly available sustainability commitments of each company with infrastructure investments on the backend, she investigated whether the data centers hosting the software workloads ran on renewable energy, at a minimum. From there, she noted whether an app's features followed green software principles [3]. She took into consideration variables such as graphics quality, "bursty" versus sustained workloads, the frequency of pings made to distant servers, and more.
After receiving input from subject matter experts, investigating industry reports, and partnering with researchers on representative workflow samples, Tristin's interpretation of Green UX produced the most comprehensive grading scheme I've yet seen. Weighting corporate sustainability commitments alongside engineering and design decisions provided a rich picture on where a company credibly stands on building low-carbon products.
Even more interesting patterns emerged when Tristin mapped her results back to the original day-in-the-life user scenarios with fellow designer Mikaela Watson. At the time of this project, remote work was firmly entrenched, with many research participants working from home. New "rush hours" for energy consumption were observable in the data at the beginning and peak of the work day.
Mornings often involved video-heavy group conference calls and multitasking across devices, whereas afternoons might be spent in collaborative workshops requiring a lot of live brainstorming on note-taking software. These large group sessions with a mix of video and interactive broadcasting made me wonder about bandwidth. Should we be comparing the energy savings of working from home—through reduced commuter traffic on roads—with the energy expenditure growing through digital networks?
In the wake of the pandemic, the adoption of hybrid work practices creates additional client-server dependencies in workplaces supporting both local and remote colleagues. And like the renewable energy consumption dilemma mentioned above, this does not equate to a reduction in work hours so much as their multiplication over virtual and real-world settings. These experiments raised many questions, which I hope will entice others to explore further.
Building Green Momentum
Nonprofits such as the Green Software Foundation (GSF; https://greensoftware.foundation/), the Green Web Foundation (GWF; https://www.thegreenwebfoundation.org/) and the Sustainable Games Alliance (SGA; https://sustainablegamesalliance.org/) are at the forefront of efforts to train developers and IT leaders about the carbon impact of the industry. GSF focuses on "reduction, not neutralization"—an important contrast to the energy targets of large corporate entities increasingly obscured by purchased offsets [4].
Complementing this work, sustainable UX practitioners (e.g., https://sustainableuxnetwork.com/ and https://sustainableux.com/) provide valuable resources, publications, and conferences to expand the range of methods that can assist in greening the user experience. Combined, these pockets of activity provide nascent standards for consumers seeking simple concepts to assess environmental merit.
Through this work—call it green UX, for want of a better option—researchers and designers can partner on the important work of adding rigor to the tech industry's sustainability goals. Connecting a product's default settings and technical affordances to the broader corporate narrative on sustainability raises awareness of the way cumulative business decisions contribute to generating overall energy demand.
To reach the end goal of carbon reduction at scale—of "computing within limits" (https://computingwithinlimits.org/2024/), as the influential scholarly conference puts it—we cannot rely solely on sustainability program managers "chasing carbon" [5], engineers optimizing code, or infrastructure teams moving server farms to green grids. All these jobs are critical to the long-term viability of the industry and should be supplements to a more commonplace literacy around resource use and computing.
Designers have a historic opportunity to get involved in sustainability education to help technology users see the effects of hardware and software choices, and the fragility of the ecosystem underpinning them.
References
1. Huang, K. Meta's AI assistant wins millions of users in challenge to ChatGPT. The Information. Aug. 29, 2024; https://www.theinformation.com/articles/metas-ai-assistant-wins-millions-of-users-in-challenge-to-chatgpt
2. Oldani, T. Applying UX methods to uncover and depict the environmental impact of digital products. Bootcamp. May 25, 2023; https://bootcamp.uxdesign.cc/applying-ux-methods-to-uncover-and-depict-the-environmental-impact-of-digital-products-d62778a1c9c4
3. Green Software Foundation. What is green software? Aug. 3, 2021; https://greensoftware.foundation/articles/what-is-green-software
4. Cockcroft, A. Sustainability: How Did Amazon, Azure, Google Perform in 2023? The New Stack. Jul. 17, 2024; https://thenewstack.io/sustainability-how-did-amazon-azure-google-perform-in-2023/
5. Gupta, U. et al. Chasing carbon: The elusive environmental footprint of computing. IEEE Micro 42, 4 (2022), 37–47.
Author
Melissa Gregg is a writer, ethnographer, and sustainability strategist to the tech industry, as well as an incoming professor of digital futures at the University of Bristol. [email protected]
Copyright held by owner/author
The Digital Library is published by the Association for Computing Machinery. Copyright © 2024 ACM, Inc.
Post Comment
No Comments Found