arXiv subjects

Maryna Sydorova

Publications and source records attributed to Maryna Sydorova.

12 recordsLinked to original sources

arXiv2026

SWARM: A Multilingual Human-Annotated Dataset for Russian Propaganda Detection in Search Engine Results

Russian state propaganda spreads across many languages and online spaces. Yet, most computational work examines only one such space, usually social media, in one or two languages, and analyses sources rather than content. We introduce SWARM (Search-Web documents Annotated for Russian propaganda, Multilingual), a dataset of 2,183 search engine results across nine languages and diverse web domains (e.g., news, blogs, government sites), each annotated by trained coders for whether it supports a recurring Russian propaganda narrative. We benchmark a source-based blocklist, supervised classifiers, and zero-shot LLMs against these labels. The blocklist misses most propaganda-supporting documents, because such content is not confined to flagged "propaganda" outlets but also appears on mainstream ones. Content-level analysis helps, though how much depends on the model: the strongest LLM reaches a positive-class F1 of 0.73, whereas the supervised classifiers reach only about 0.5, with the smaller LLMs over-predicting support, mistaking topical relevance for endorsement. Detecting search-borne propaganda thus requires per-language, content-level evaluation, which we hope SWARM and our evaluation code enable.

cs.CL↗

arXiv2026

High-Risk Memories? Comparative audit of the representation of Second World War atrocities in Ukraine by generative AI applications

The rise of generative artificial intelligence (genAI) models poses new possibilities and risks for how the past is remembered by accelerating content production and altering the process of information discovery. The most critical risk is historical misrepresentation, which ranges from the distortion of facts and inaccurate depiction of specific groups to more subtle forms, such as the selective moralization of history. The dangers of misrepresentation of the past are particularly pronounced for high-risk memories, such as memories of past atrocities, which have a strong emotional load and are often instrumentalised by political actors. To understand how substantive this risk is, we empirically investigate how genAI applications deal with high-risk memories of the Second World War atrocities in Ukraine. This case is crucial due to the scope of the atrocities and the intense, often instrumentalised, contestation surrounding their memory. We audit the performance of three common genAI applications for different types of misrepresentation, including hallucinations and inconsistent moralization, and discuss the implications for future memory practices.

cs.CY↗

arXiv2025

Invisible in Search? Auditing Aesthetic Bias in the Visual Representation of Holocaust Victims on Google

Information retrieval systems, such as search engines, increasingly shape the representation of the past and present states of social reality. Despite their importance, these systems face challenges in dealing with the ethical aspects of representation due to various forms of bias, including aesthetic bias that perpetuates hegemonic patterns of representation. While most research on aesthetic bias has examined it in the context of current societal issues, it is also crucial for historical representation, particularly of sensitive subjects such as historical atrocities. To address this gap, we conduct a comparative audit of the visual representation of Holocaust victims on Google. We find that Google tends to propagate a male-dominated representation of Holocaust victims with an emphasis on atrocity context, risking rendering invisible gender-specific suffering and decreasing potential for nurturing empathy. We also observe a variation in representation across geographic locations, suggesting that search algorithms may produce their own aesthetic of victimhood.

cs.CY↗

arXiv2025

Campaigning through the lens of Google: A large-scale algorithm audit of Google searches in the run-up to the Swiss Federal Elections 2023

Search engines like Google have become major sources of information for voters during election campaigns. To assess potential biases across candidates' gender and partisan identities in the algorithmic curation of candidate information, we conducted a large-scale algorithm audit analyzing Google's selection and ranking of information about candidates for the 2023 Swiss Federal Elections, three and one week before the election day. Results indicate that text searches prioritize media sources in search output but less so for women politicians. Image searches revealed a tendency to reinforce stereotypes about women candidates, marked by a disproportionate focus on stereotypically pleasant emotions for women, particularly among right-leaning candidates. Crucially, we find that patterns of candidates' representation in Google text and image searches are predictive of their electoral performance.

cs.CY↗

arXiv2025

Search engines in polarized media environment: Auditing political information curation on Google and Bing prior to 2024 US elections

Search engines play an important role in the context of modern elections. By curating information in response to user queries, search engines influence how individuals are informed about election-related developments and perceive the media environment in which elections take place. It has particular implications for (perceived) polarization, especially if search engines' curation results in a skewed treatment of information sources based on their political leaning. Until now, however, it is unclear whether such a partisan gap emerges through information curation on search engines and what user- and system-side factors affect it. To address this shortcoming, we audit the two largest Western search engines, Google and Bing, prior to the 2024 US presidential elections and examine how these search engines' organic search results and additional interface elements represent election-related information depending on the queries' slant, user location, and time when the search was conducted. Our findings indicate that both search engines tend to prioritize left-leaning media sources, with the exact scope of search results' ideological slant varying between Democrat- and Republican-focused queries. We also observe limited effects of location- and time-based factors on organic search results, whereas results for additional interface elements were more volatile over time and specific US states. Together, our observations highlight that search engines' information curation actively mirrors the partisan divides present in the US media environments and has the potential to contribute to (perceived) polarization within these environments.

cs.CY↗

arXiv2025

Examining bias perpetuation in academic search engines: an algorithm audit of Google and Semantic Scholar

Researchers rely on academic Web search engines to find scientific sources, but search engine mechanisms may selectively present content that aligns with biases embedded in queries. This study examines whether confirmation biased queries prompted into Google Scholar and Semantic Scholar will yield results aligned with a query's bias. Six queries (topics across health and technology domains such as vaccines, Internet use) were analyzed for disparities in search results. We confirm that biased queries (targeting benefits or risks) affect search results in line with bias, with technology-related queries displaying more significant disparities. Overall, Semantic Scholar exhibited fewer disparities than Google Scholar. Topics rated as more polarizing did not consistently show more disparate results. Academic search results that perpetuate confirmation bias have strong implications for both researchers and citizens searching for evidence. More research is needed to explore how scientific inquiry and academic search engines interact.

cs.CY↗

arXiv2024

Google, How Should I Vote? How Users Formulate Search Queries to Find Political Information on Search Engines

Search engine results depend not only on the algorithms but also on how users interact with them. However, factors affecting the selection of a search query remain understudied. Using a representative survey of Swiss citizens before a round of federal popular votes, this study examines how users formulate search queries related to the retirement policies that were voted on in March 2024. Contrary to existing research, we find no direct evidence of selective exposure, or users' tendency to search for pro-attitudinal information, which we explain by the less polarizing search topics. However, we find that the sentiment of the query is partially aligned with the expected vote outcome. Our results also suggest that undecided and non-voters are more likely to search for nuanced information, such as consequences and interpretations of the policies. The perceived importance and effect of the issue, political views, and sociodemographics also affect query formulation.

cs.HC↗

arXiv2024

LLMs as information warriors? Auditing how LLM-powered chatbots tackle disinformation about Russia's war in Ukraine

The rise of large language models (LLMs) has a significant impact on information warfare. By facilitating the production of content related to disinformation and propaganda campaigns, LLMs can amplify different types of information operations and mislead online users. In our study, we empirically investigate how LLM-powered chatbots, developed by Google, Microsoft, and Perplexity, handle disinformation about Russia's war in Ukraine and whether the chatbots' ability to provide accurate information on the topic varies across languages and over time. Our findings indicate that while for some chatbots (Perplexity), there is a significant improvement in performance over time in several languages, for others (Gemini), the performance improves only in English but deteriorates in low-resource languages.

cs.CY↗

arXiv2024

Finding frames with BERT: A transformer-based approach to generic news frame detection

Framing is among the most extensively used concepts in the field of communication science. The availability of digital data offers new possibilities for studying how specific aspects of social reality are made more salient in online communication but also raises challenges related to the scaling of framing analysis and its adoption to new research areas (e.g. studying the impact of artificial intelligence-powered systems on representation of societally relevant issues). To address these challenges, we introduce a transformer-based approach for generic news frame detection in Anglophone online content. While doing so, we discuss the composition of the training and test datasets, the model architecture, and the validation of the approach and reflect on the possibilities and limitations of the automated detection of generic news frames.

cs.CL↗

arXiv2024

Finding the white male: The prevalence and consequences of algorithmic gender and race bias in political Google searches

Search engines like Google have become major information gatekeepers that use artificial intelligence (AI) to determine who and what voters find when searching for political information. This article proposes and tests a framework of algorithmic representation of minoritized groups in a series of four studies. First, two algorithm audits of political image searches delineate how search engines reflect and uphold structural inequalities by under- and misrepresenting women and non-white politicians. Second, two online experiments show that these biases in algorithmic representation in turn distort perceptions of the political reality and actively reinforce a white and masculinized view of politics. Together, the results have substantive implications for the scientific understanding of how AI technology amplifies biases in political perceptions and decision-making. The article contributes to ongoing public debates and cross-disciplinary research on algorithmic fairness and injustice.

cs.CY↗

arXiv2024

Algorithmically Curated Lies: How Search Engines Handle Misinformation about US Biolabs in Ukraine

The growing volume of online content prompts the need for adopting algorithmic systems of information curation. These systems range from web search engines to recommender systems and are integral for helping users stay informed about important societal developments. However, unlike journalistic editing the algorithmic information curation systems (AICSs) are known to be subject to different forms of malperformance which make them vulnerable to possible manipulation. The risk of manipulation is particularly prominent in the case when AICSs have to deal with information about false claims that underpin propaganda campaigns of authoritarian regimes. Using as a case study of the Russian disinformation campaign concerning the US biolabs in Ukraine, we investigate how one of the most commonly used forms of AICSs - i.e. web search engines - curate misinformation-related content. For this aim, we conduct virtual agent-based algorithm audits of Google, Bing, and Yandex search outputs in June 2022. Our findings highlight the troubling performance of search engines. Even though some search engines, like Google, were less likely to return misinformation results, across all languages and locations, the three search engines still mentioned or promoted a considerable share of false content (33% on Google; 44% on Bing, and 70% on Yandex). We also find significant disparities in misinformation exposure based on the language of search, with all search engines presenting a higher number of false stories in Russian. Location matters as well with users from Germany being more likely to be exposed to search results promoting false information. These observations stress the possibility of AICSs being vulnerable to manipulation, in particular in the case of the unfolding propaganda campaigns, and underline the importance of monitoring performance of these systems to prevent it.

cs.IR↗

arXiv2022

Panning for gold: Lessons learned from the platform-agnostic automated detection of political content in textual data

The growing availability of data about online information behaviour enables new possibilities for political communication research. However, the volume and variety of these data makes them difficult to analyse and prompts the need for developing automated content approaches relying on a broad range of natural language processing techniques (e.g. machine learning- or neural network-based ones). In this paper, we discuss how these techniques can be used to detect political content across different platforms. Using three validation datasets, which include a variety of political and non-political textual documents from online platforms, we systematically compare the performance of three groups of detection techniques relying on dictionaries, supervised machine learning, or neural networks. We also examine the impact of different modes of data preprocessing (e.g. stemming and stopword removal) on the low-cost implementations of these techniques using a large set (n = 66) of detection models. Our results show the limited impact of preprocessing on model performance, with the best results for less noisy data being achieved by neural network- and machine-learning-based models, in contrast to the more robust performance of dictionary-based models on noisy data.

cs.CL↗