Where the Earth is flat and 9/11 is an inside job: A comparative algorithm audit of conspiratorial information in web search results

Different platforms, different plots? The Kremlin-controlled search engine Yandex as a resource for Russia’s informational influence in Belarus during the COVID-19 pandemic

“…The shares of conspiratorial content on Bing, DuckDuckGo and Yahoo varied between 25% and 30%. Similar to Hussein et al (2020) , Urman et al (2021b) found that the proportions of the retrieved CTs varied between queries and suggested that Google was modifying its search results to exclude conspiratorial content. They further suggested that the content differences between Google and Yandex with regard to CTs were partly due to the differences in the source types featured by the engines: while Google had the highest share of scientific sources, those were absent on Yandex.…”

“…In addition, our findings demonstrated profound differences with regard to the prioritisation of conspiratorial content between Yandex and Google: Yandex’s web search algorithms were more than two times more likely to retrieve undebunked CTs in its search results than Google’s (H1). While extant research has typically suggested that the content differences between Google and Yandex with regard to CTs are mostly due to the differences in the prioritised source types ( Urman et al, 2021b ), moderation policies ( Hussein et al, 2020 ; Urman et al, 2021b ) or political alignment of the sources of the search results ( Makhortykh et al, 2022 ), this study was the first to consider the link between the geographic origin of the sources of search results (Russia-based vs. non-Russia-based) and the extent of the conspiratorial content present in the search results. In the context of this study, we found that the geographic origin of the search results played a major role (RQ1/RQ3).…”

“…Considering the importance of search engines in the digitalised news-gathering process ( Fletcher and Nielsen, 2018 ; Wallace, 2018 ), the presence of CTs in top search results could induce the formation of conspiracy beliefs. Yet, the role of search engines in promoting CTs has received little scholarly attention (for similar claims see Bandy, 2021 ), with a few notable exceptions in recent years ( Hussein et al, 2020 ; Toepfl et al, 2022 , 2023 ; Urman et al, 2021b ). For instance, Hussein et al (2020) investigated how YouTube’s search engine promotes misinformative content.…”

“…They further suggested that the content differences between Google and Yandex with regard to CTs were partly due to the differences in the source types featured by the engines: while Google had the highest share of scientific sources, those were absent on Yandex. Instead, Yandex prioritised ‘conspiracy-dedicated’ ( Urman et al, 2021b : 14) websites and links to social media.…”

“…Extant research has found Yandex’s web search algorithms, compared with Google’s, to systematically favor conspiratorial content in its search results ( Makhortykh et al, 2022 ; Urman et al, 2021b ). Against this backdrop, we hypothesise that COVID-19-related CTs are more visible in Yandex’s web search results also from Belarus:…”