Search

Search engines, such as Google or Yandex, shape social reality by informing their users about current and historical phenomena. However, there is little research on how search engines deal with contested memories, which are subjected to ontological conflicts known as memory wars. In this article, we investigate how search engines circulate information about memory wars related to the Holodomor, a mass famine caused by Soviet repressive politics in Ukraine in 1932–1933. For this aim, we conduct an agent-based audit of four search engines—Bing, DuckDuckGo, Google, and Yandex—and examine how their top search results represent the Holodomor and related memory wars. Our findings demonstrate that search engines prioritize interpretations of the Holodomor aligning with specific sides in the memory wars, thus becoming memory warriors themselves.

Search engines play a gatekeeper role in current high-choice information environments. Considered a form of new media, users are still more likely to find and trust news found through search than social media sites. Indeed, search engines are one of the most utilised technologies to find political information, despite audits uncovering biases in their results, for example, towards national outlets over local ones. It is therefore important to keep in mind the potential of search results to affect public opinion. With this study, we investigate how Google search news headlines and snippets differ when varying migrant search terms (e.g., immigrant, refugee, expat). We employ computational text analysis methods as well as qualitative content analysis. Specifically, we employ an automated framework for detecting media frames, originally trained on Twitter data, and attempt to transfer it to news data; this framework allows for a categorization of data to frames of a generic-issue (economy, safety, health) and specific (hero:diversity, threat:jobs) nature. We evaluate its applicability for this novel data source and find that it performs well for frames related economy and security. Our next steps include analysing the results of other computational measures, namely, sentiment, agency and political outlet of the news item. We expect that sentiment and agency will complement the initial results we see based on media frames.

Search engines serve as information gatekeepers on a multitude of topics dealing with different aspects of society. However, the ways search engines filter and rank information are prone to biases related to gender, ethnicity, and race. In this paper, we conduct a systematic algorithm audit to examine how one specific form of misrepresentation bias, namely, sexualization, is manifested in Google's text search results about different national and gender groups. We find evidence of sexualization of women, particularly those from the Global South and East in search outputs. This is amplified by contextual advertisements pushed by Google in response to different search queries.

On 24 February 2022, Russia launched a large-scale invasion of Ukraine which since then has been consistently referred to by the Russian authorities as a special operation and not war. The purpose of this specific framing of the Russian aggression can be attributed to multiple reasons, but among the main advantages it provides to the Kremlin is the emphasis of power inequality between Russia and Ukraine as well as minimizing the concerns of the Russian population about the impact of the war on their lives. However, despite all the Kremlin’s propaganda and censorship efforts, it is still unclear what the dominant framing of the war is among regular Russian internet users and how they engage with different frames used in relation to it. To address this gap, we conducted an automated analysis of over 6 million public posts from VK, the biggest Russian social media platform, published between 14 February 2022 and 24 June 2022 and related to different ways of framing the war as well as the ways users engage with these frames. Our analysis reveals that the invasion was predominantly discussed in connection with the losses in the Russian army, and allusions to WW2 were also common. The users who posted about the war were older than average VK users and tuned in to pro-regime and pro-war content more frequently.

The present paper contributes to the research on the activities of far-right actors on social media by examining the interconnections between far-right actors and groups on Telegram platform using ...

We examine the patterns of medical preprint sharing on Twitter during the early stages of the COVID-19 pandemic. Our analysis demonstrates a stark increase in attention to medical preprints among the general public since the beginning of the pandemic. We also observe a political divide in medical preprint sharing patterns - a finding in line with previous observations regarding the politicisation of COVID-19-related discussions. In addition, we find that the increase in attention to preprints from the members of the general public coincided with the change in the social media-based discourse around preprints.

PurposeThis study investigates perceptions of the use of online tracking, a passive data collection method relying on the automated recording of participant actions on desktop and mobile devices, for studying information behavior. It scrutinizes folk theories of tracking, the concerns tracking raises among the potential participants and design mechanisms that can be used to alleviate these concerns.Design/methodology/approachThis study uses focus groups composed of university students (n = 13) to conduct an in-depth investigation of tracking perceptions in the context of information behavior research. Each focus group addresses three thematic blocks: (1) views on online tracking as a research technique, (2) concerns that influence participants' willingness to be tracked and (3) design mechanisms via which tracking-related concerns can be alleviated. To facilitate the discussion, each focus group combines open questions with card-sorting tasks. The results are analyzed using a combination of deductive content analysis and constant comparison analysis, with the main coding categories corresponding to the thematic blocks listed above.FindingsThe study finds that perceptions of tracking are influenced by recent data-related scandals (e.g. Cambridge Analytica), which have amplified negative attitudes toward tracking, which is viewed as a surveillance tool used by corporations and governments. This study also confirms the contextual nature of tracking-related concerns, which vary depending on the activities and content that are tracked. In terms of mechanisms used to address these concerns, this study highlights the importance of transparency-based mechanisms, particularly explanations dealing with the aims and methods of data collection, followed by privacy- and control-based mechanisms.Originality/valueThe study conducts a detailed examination of tracking perceptions and discusses how this research method can be used to increase engagement and empower participants involved in information behavior research.

The growth of online platforms is accompanied by the increasing use of automated agents. Despite being discussed primarily in the context of opinion manipulation, agents play diverse roles within platform ecosystems that raises the need for governance approaches that go beyond policing agents’ unwanted behaviour. To provide a more nuanced assessment of agent governance, we introduce an analytical framework that distinguishes between different aspects and forms of governance. We then apply it to explore how agents are governed across nine platforms. Our observations show that despite acknowledging diverse roles of agents, platforms tend to focus on governing selected forms of their misuse. We also observe differences in governance approaches used by platforms, in particular when it comes to the agent rights/obligations and transparency of policing mechanisms. These observations highlight the necessity of advancing the algorithmic governance research agenda and developing a generalizable normative framework for agent governance.

Implicit and explicit gender biases in media representations of individuals have long existed. Women are less likely to be represented in gender-neutral media content (representation bias), and their face-to-body ratio in images is often lower (face-ism bias). In this article, we look at representativeness and face-ism in search engine image results. We systematically queried four search engines (Google, Bing, Baidu, Yandex) from three locations, using two browsers and in two waves, with gender-neutral (person, intelligent person) and gendered (woman, intelligent woman, man, intelligent man) terminology, accessing the top 100 image results. We employed automatic identification for the individual’s gender expression (female/male) and the calculation of the face-to-body ratio of individuals depicted. We find that, as in other forms of media, search engine images perpetuate biases to the detriment of women, confirming the existence of the representation and face-ism biases. In-depth algorithmic debiasing with a specific focus on gender bias is overdue.

By filtering and ranking information, search engines shape how individuals perceive both the present and past events. However, these information curation mechanisms are prone to malperformance that can misinform their users. In this article, we examine how search malperformance can influence representation of traumatic past by investigating image search outputs of six search engines in relation to the Holocaust in English and Russian. Our findings indicate that besides two common themes - commemoration and liberation of camps - there is substantial variation in visual representation of the Holocaust between search engines and languages. We also observe several instances of search malperformance, including content propagating antisemitism and Holocaust denial, misattributed images, and disproportionate visibility of specific Holocaust aspects that might result in its distorted perception by the public.

We audit the presence of domain-level source diversity bias in video search results. Using a virtual agent-based approach, we compare outputs of four Western and one non-Western search engines for English and Russian queries. Our findings highlight that source diversity varies substantially depending on the language with English queries returning more diverse outputs. We also find disproportionately high presence of a single platform, YouTube, in top search outputs for all Western search engines except Google. At the same time, we observe that Youtube’s major competitors such as Vimeo or Dailymotion do not appear in the sampled Google’s video search results. This finding suggests that Google might be downgrading the results from the main competitors of Google-owned Youtube and highlights the necessity for further studies focusing on the presence of own-content bias in Google’s search results.

Web search engines influence perception of social reality by filtering and ranking information. However, their outputs are often subjected to bias that can lead to skewed representation of subjects such as professional occupations or gender. In our paper, we use a mixed-method approach to investigate presence of race and gender bias in representation of artificial intelligence (AI) in image search results coming from six different search engines. Our findings show that search engines prioritize anthropomorphic images of AI that portray it as white, whereas non-white images of AI are present only in non-Western search engines. By contrast, gender representation of AI is more diverse and less skewed towards a specific gender that can be attributed to higher awareness about gender bias in search outputs. Our observations indicate both the the need and the possibility for addressing bias in representation of societally relevant subjects, such as technological innovation, and emphasize the importance of designing new approaches for detecting bias in information retrieval systems., Comment: 16 pages, 3 figures

In this article, we conduct a comparative analysis of web search behaviors in Switzerland and Germany. For this aim, we rely on a combination of web tracking data and survey data collected over a period of 2 months from users in Germany (n = 558) and Switzerland (n = 563). We find that web search accounts for 13% of all desktop browsing, with the share being higher in Switzerland than in Germany. In over 50% of cases users clicked on the first search result, with over 97% of all clicks being made on the first page of search outputs. Most users rely on Google when conducting searches, with some differences observed in users’ preferences for other engines across demographic groups. Further, we observe differences in the temporal patterns of web search use between women and men, marking the necessity of disaggregating data by gender in observational studies regarding online information seeking behaviors. Our findings highlight the contextual differences in web search behavior across countries and demographic groups that should be taken into account when examining search behavior and the potential effects of web search result quality on societies and individuals.

BACKGROUND Since the beginning of the COVID-19 pandemic, scientists have been willing to share their results quickly to speed up the development of potential treatments and/or a vaccine. At the same time, traditional peer-review-based publication systems are not always able to process new research promptly. This has contributed to a surge in the number of medical preprints published since January 2020. In the absence of a vaccine, preventative measures such as social distancing are most helpful in slowing the spread of COVID-19. Their effectiveness can be undermined if the public does not comply with them. Hence, public discourse can have a direct effect on the progression of the pandemic. Research shows that social media discussions on COVID-19 are driven mainly by the findings from preprints, not peer-reviewed papers, highlighting the need to examine the ways medical preprints are shared and discussed online. OBJECTIVE We examine the patterns of medRxiv preprint sharing on Twitter to establish (1) whether the number of tweets linking to medRxiv increased with the advent of the COVID-19 pandemic; (2) which medical preprints were mentioned on Twitter most often; (3) whether medRxiv sharing patterns on Twitter exhibit political partisanship; (4) whether the discourse surrounding medical preprints among Twitter users has changed throughout the pandemic. METHODS The analysis is based on tweets (n=557,405) containing links to medRxriv preprint repository that were posted between the creation of the repository in June 2019 and June 2020. The study relies on a combination of statistical techniques and text analysis methods. RESULTS Since January 2020, the number of tweets linking to medRxiv has increased drastically, peaking in April 2020 with a subsequent cool-down. Before the pandemic, preprints were shared predominantly by users we identify as medical professionals and scientists. After January 2020, other users, including politically-engaged ones, have started increasingly tweeting about medRxiv. Our findings indicate a political divide in sharing patterns of the top-10 most-tweeted preprints. All of them were shared more frequently by users who describe themselves as Republicans than by users who describe themselves as Democrats. Finally, we observe a change in the discourse around medRxiv preprints. Pre-pandemic tweets linking to them were predominantly using the word “preprint”. In February 2020 “preprint” was taken over by the word “study”. Our analysis suggests this change is at least partially driven by politically-engaged users. Widely shared medical preprints can have a direct effect on the public discourse around COVID-19, which in turn can affect the societies’ willingness to comply with preventative measures. This calls for an increased responsibility when dealing with medical preprints from all parties involved: scientists, preprint repositories, media, politicians, and social media companies. CONCLUSIONS Widely shared medical preprints can have a direct effect on the public discourse around COVID-19, which in turn can affect the societies’ willingness to comply with preventative measures. This calls for an increased responsibility when dealing with medical preprints from all parties involved: scientists, preprint repositories, media, politicians, and social media companies.

In this paper, we examine the effects of the YouTube recommendation algorithm on the distribution of ideologically-charged news content. For that purpose, we develop a research infrastructure and conduct a series of experiments using virtual agents (n=200) in a fully controlled environment. We specifically look at YouTube recommendations for videos related to the far-right terrorist attack in the German city of Halle in 2019 and examine how these recommendations differ depending on the type and political affiliation of videos watched by the agents. We find that YouTube recommendations are highly randomized that leads to fundamentally different recommendation trajectories under the condition of identical agent activity which was also synchronized to isolate the effect of time. We also find significant discrepancies in recommendations generated in browsers, with the recommendations for Firefox being slightly less randomized than those for Chrome. Finally, our observations suggest that the recommendations for the agents starting with right-leaning news videos are marginally more consistent than those for the mainstream and left-leaning videos.