Almost everyone thinks they know how to use Google and they usually get the answer they want. Most people will intuitively know that the query “milk good for you” leads to different results than “milk bad for you”. The same goes for “climate change” or “climate fraud” or “the 2020 US Election is valid” and “stop stealing” requests.
Because search engines are more of a “wish list” than an authoritative source, they can help spread false and misinformation that can be harmful to democracy or society. They are not neutral data brokers.
Instead, search engines return a list of results they think are most relevant to a particular query. Core algorithms make decisions about relevance and visibility for a specific query in a specific location and sometimes for a specific user.
Search engines are an integral but often invisible part of how people navigate today’s world. In this function, they also shape their understanding of reality and may thereby harm the environment. In a recent paper, we argue that the assumptions that search engines make about what we’re looking for can cause people to emit more carbon than they would otherwise.
The environmental damage of algorithmic curation
Take the “summer dress” query as an example. You will receive a list of online or nearby stores that sell summer dresses, as well as pictures of fashion models displaying the dresses for sale. This is what we are waiting for.
GoogleCC BY-SA
However, other possible interpretations of the “summer dress” query are possible. Perhaps you would like to find out what summer dresses were like in a certain historical period. Maybe you want to see what colors are best to wear in your wardrobe this year. Or maybe you really want to buy summer clothes, but only from organic or fair trade certified fabrics or from a thrift store.
You can also enter the names of two major cities, such as “Berlin Stockholm”. Google will provide you with results mainly related to air travel, but not a comparison of the livability of these cities, for example. Google will highlight different flight options in the domestic flight comparison, and you have to scroll down to find train tickets.
These results are by no means predetermined, but rather the result of algorithmic curation. Even without personalization, search results listings are uniquely created from specific searches, search engine algorithms, and specific content optimized for the user’s query and location.
You can try it yourself in these and other cities. But note that the use of quotation marks, the order of cities, or the use of local and English spelling can make a difference, as many companies try to optimize for specific searches.
Any reader who is familiar with Google search knows that these alternative results that we mention must require other queries. Such requests should clearly state that the search is about something other than buying clothes or a flight. For example, with the query “colors of summer clothes” or “Berlin is livable in Stockholm”.
Either way, the default choices that algorithms select and curate shape what we think of as default. If we are not careful and thoughtful about our goals when searching, it will affect the actions of at least some people. And these actions have very real environmental impacts.
Environmental damage as algorithmic damage
We propose to call these environmental impacts “algorithmically embodied emissions”. By this, we mean potential emissions in content that search engines or algorithmic information systems such as Facebook or TikTok feeds offer as a default option.
Our work so far is conceptual, although we hope to develop a way to quantify the concept in the future. For now, we can observe that the search results offer high-carbon practices.
And we can mention that related companies such as flight comparison services or fast fashion brands can also optimize their websites for better search engine rankings. These companies tend to have larger budgets than their more sustainable alternatives (such as a small organic or recycled summer clothing brand).
In recent years, researchers have highlighted the potential harm that algorithmic decision-making can do to people, such as replicating racial or gender biases. This is often called algorithmic damage.
The concept of algorithmically embodied emissions requires us to take algorithmic harms even further. It shows that algorithmic decision-making has a real impact on both people and the planet.
It’s also an example of how algorithmic decision-making has higher-level effects than the immediate harm it causes to individuals. In other words: how algorithms work and shape our actions matters. Even as the climate crisis accelerates, we are only just beginning to question how algorithms shape our thoughts and actions about the environment.
In response to this article, a Google spokesperson said:
At Google Search, our goal is to connect people with timely, relevant, and useful information to make sustainable choices easier. We fundamentally design our search ranking systems to uncover high-quality, reliable information on topics like climate change. To complement these efforts, we’ve developed a number of features to help people quickly access information about the environmental impact of the goods and services they see in the results, giving them useful context for making informed decisions about sustainability. We collaborate with thousands of partners across industries, from cities and governments to companies and nonprofits, to drive sustainability and climate progress.