Do Google Search Recommendations Influence the Public Debate?

The next session at AoIR 2015 is exploring timing issues, and the first paper by Sarah Muñoz-Bates is about the effects of Google on how people perceive topics. For example, what is the effect of seeing the term 'illegal' rather than 'undocumented' in relation to migrants? Does it cross the line and criminalise the person; is it racialised in a way that other terms are not?

Google participates in this debate with the word choices that it provides to users as they enter their search queries. These are largely driven by the words and phrases that Web users and content creators themselves are using – or at least that is the explanation Google itself has been using whenever this has been raised in controversial cases –, but at the same time Google's algorithms are shaping these selections as well.

This is not a question about who is in control, but about the interactions between Google and its users as topics are shaped. This shaping is dynamic, too, and recommended search terms change over time. Sarah conducted her research by exploring the search autocompletes and recommendations for the term 'immigration', compared across US, Mexico, and Dominican Republic Google sites.

This yielded a diverse range of related topics, including terms related to illegal, clandestine, and undocumented immigration; in the US, 'undocumented' remained absent from search recommendations. This has slowly changed in more recent time. In the other countries, words such as 'clandestine' and 'irregular', as well as terms relating to xenophobia appeared more prominently. This is not consciously driven by Google, Sarah suggests, but the words that Google provides, and the way that they change over time, can influence the public debate.