arXiv subjects

Kiran Garimella

Publications and source records attributed to Kiran Garimella.

At least 19 recordsLinked to original sources

arXiv2026

How Students (Really) Use ChatGPT: Uncovering Experiences Among Undergraduate Students

We examine how undergraduate students integrate ChatGPT into everyday self-directed learning, analyzing 10,536 naturalistic messages donated by 36 students over a year. A sequential mixed-methods pipeline pairs iterative qualitative coding with zero-shot language-model annotation validated against human labels (kappa = 0.75-0.91). It yields a five-category taxonomy: Information Seeking, Content Generation, Language Use, Student-ChatGPT Interaction, and ChatGPT Response Behavior. Time-lagged linear regression and Cox proportional-hazards models link these categories to sustained engagement. Three findings stand out. First, structured tasks (theory application, code writing, job-application content, multiple-choice questions) predict continued use; ChatGPT becomes incorporated into academic rhythms when gratifications are reliably fulfilled. Second, system-issued "apologies" are the strongest positive predictor of increased engagement, outweighing every task-completion predictor. We name this mechanism "repair gratification": the reward of a breakdown acknowledged and repaired rather than a task simply completed. Third, interactional strain--prompt revision, frustration, follow-up clarification--predicts disengagement. When managing the system falls on the user without system accountability, students abandon the tool. We interpret these results through Self-Directed Learning, Uses and Gratifications Theory, and Expectancy Violations Theory, mapping predictors onto positive/negative violations and confirmations. We close with design recommendations for graduated repair patterns, mode-aware interaction, and verification affordances, and outline a participatory AI-literacy agenda for higher education.

cs.HC↗

arXiv2026

From Retrieval to Synthesis: Repair Literacy and the Domestication of Generative AI

How do students develop AI literacy through everyday practice rather than formal instruction? While normative AI literacy frameworks proliferate, empirical understanding of how students actually learn to work with generative AI remains limited. This study analyzes 10,536 ChatGPT messages from 36 undergraduates over one academic year, revealing five use genres -- academic workhorse, emotional companion, metacognitive partner, repair and negotiation, and trust calibration -- that constitute distinct configurations of student-AI learning. Drawing on domestication theory and emerging frameworks for AI literacy, we demonstrate that functional AI competence emerges through ongoing relational negotiation rather than one-time adoption. Students develop sophisticated genre portfolios, strategically matching interaction patterns to learning needs while exercising critical judgment about AI limitations. Notably, repair work during AI breakdowns produces substantial learning about AI capabilities, developing what we term "repair literacy" -- a crucial but underexplored dimension of AI competence. Our findings offer educators empirically grounded insights into how students actually learn to work with generative AI, with implications for AI literacy pedagogy, responsible AI integration, and the design of AI-enabled learning environments that support student agency.

cs.HC↗

arXiv2026

Creating Group Rules with AI: Human-AI Collaboration in WhatsApp Moderation

WhatsApp is one of the most widely used messaging platforms globally, with billions of users sharing information in private groups. Yet, it offers little infrastructure to support moderation and group governance. In the absence of platform-level oversight, group admins bear the responsibility of governing group behavior. In this paper, we explore how WhatsApp group admins collaborate with AI tools to create, enforce, and maintain group rules. Drawing on a two-phase speculative design study with 20 admins in India, we examine how participants interacted with an AI assistant (Meta AI) to co-create rules and responded to a series of probes illustrating AI-assisted moderation features. Our findings show that while admins appreciated the AI's ability to surface overlooked rules and reduce their moderation burden, they were highly sensitive to issues of relational trust, data privacy, tone, and social context. We identify how group type and admin style shaped their willingness to delegate authority, and surface the limitations of current chatbot interfaces in supporting collaborative rule-making. We conclude with design implications for building moderation tools that center human judgment, relational nuance, contextual adaptability, and collective governance.

cs.HC↗

arXiv2026

Inferential Privacy Leakage in Anonymized Conversational AI Logs

Hundreds of millions of users now hold detailed, multi-turn conversations with ChatGPT and similar LLM assistants. We measure two privacy-relevant features of these conversations on a corpus of complete ChatGPT histories donated by over 1,000 users in four Global South countries (Brazil, India, Nigeria, Pakistan). First, on explicit disclosure: 34.5% of user messages contain personal information across a twenty-category taxonomy, with the median user first revealing identifying content within the first 14% of their conversation history. Second, on inference beyond explicit disclosure: we restrict to a cohort whose conversations contain no messages flagged by an LLM-based filter for explicit demographic self-identification (a separate NER pass marks PII for the disclosure audit but does not drive cohort exclusion). On this filtered cohort, an off the shelf large language model still recovers each user's age, gender, and country at weighted F1 of 0.84, 0.90, and 0.88, respectively, with the median user identified from the first 5% of their conversation history. Reading the model's natural-language reasoning traces, we identify four recurring stereotype patterns that drive both successful inference and an asymmetric error distribution concentrating on women in technical fields, older users with contemporary skills, and Global South tech professionals. We also compare ChatGPT against the same users' Google Search and YouTube histories as inference surfaces, and find it competitive with these older substrates that have driven behavioral advertising for two decades. Message-level PII removal is insufficient on its own as a privacy intervention for conversational AI data.

cs.CY↗

arXiv2026

Demographic Divides in Political Content Exposure on Facebook

Despite Facebook's central role in American civic life, a clear, evidence-based understanding of users' long-term information environments has remained elusive, hindering assessments of the platform's societal impact. This study addresses that gap by analyzing a unique decade-long dataset, constructed by collecting the full list of public pages and groups followed by over 1,100 American users. This approach allows us to examine the potential information exposure of these users by analyzing hundreds of millions of posts from 2012 to 2023. We find that political content constitutes a modest 18% of a user's potential information diet, which is predominantly composed of lifestyle and entertainment topics. This aggregate view, however, masks a deeply stratified reality: we uncover significant and persistent disparities in the volume and ideological leaning of political content across age, gender, and racial lines. Furthermore, we quantify the porous boundaries between content categories, showing how political discourse frequently permeates non-political spaces. Leveraging the dataset's longitudinal nature, we also assess the impact of major platform interventions. We find that Meta's 2018 "Meaningful Social Interactions" update dramatically increased the share of political content by contracting the visibility of non-political posts. By providing a granular, decade-long map of potential information exposure, our study offers one of the first representative and longitudinal picture drawn from platform-independent data. Our findings underscore the critical need for researchers to measure exposure, not merely engagement, and to account for the significant volume of political content that circulates in non-political spaces.

cs.SI↗

arXiv2025

Listening Between the Lines: Decoding Podcast Narratives with Language Modeling

Podcasts have become a central arena for shaping public opinion, making them a vital source for understanding contemporary discourse. Their typically unscripted, multi-themed, and conversational style offers a rich but complex form of data. To analyze how podcasts persuade and inform, we must examine their narrative structures -- specifically, the narrative frames they employ. The fluid and conversational nature of podcasts presents a significant challenge for automated analysis. We show that existing large language models, typically trained on more structured text such as news articles, struggle to capture the subtle cues that human listeners rely on to identify narrative frames. As a result, current approaches fall short of accurately analyzing podcast narratives at scale. To solve this, we develop and evaluate a fine-tuned BERT model that explicitly links narrative frames to specific entities mentioned in the conversation, effectively grounding the abstract frame in concrete details. Our approach then uses these granular frame labels and correlates them with high-level topics to reveal broader discourse trends. The primary contributions of this paper are: (i) a novel frame-labeling methodology that more closely aligns with human judgment for messy, conversational data, and (ii) a new analysis that uncovers the systematic relationship between what is being discussed (the topic) and how it is being presented (the frame), offering a more robust framework for studying influence in digital media.

cs.CL↗

arXiv2025

Source Coverage and Citation Bias in LLM-based vs. Traditional Search Engines

LLM-based Search Engines (LLM-SEs) introduces a new paradigm for information seeking. Unlike Traditional Search Engines (TSEs) (e.g., Google), these systems summarize results, often providing limited citation transparency. The implications of this shift remain largely unexplored, yet raises key questions regarding trust and transparency. In this paper, we present a large-scale empirical study of LLM-SEs, analyzing 55,936 queries and the corresponding search results across six LLM-SEs and two TSEs. We confirm that LLM-SEs cites domain resources with greater diversity than TSEs. Indeed, 37% of domains are unique to LLM-SEs. However, certain risks still persist: LLM-SEs do not outperform TSEs in credibility, political neutrality and safety metrics. Finally, to understand the selection criteria of LLM-SEs, we perform a feature-based analysis to identify key factors influencing source choice. Our findings provide actionable insights for end users, website owners, and developers.

cs.CL↗

arXiv2025

Disagreement is Disappearing on U.S. Cable Debate Shows

Prime-time cable news programs are a highly influential part of the American media landscape, with top-rated opinion shows attracting millions of politically attentive viewers each night. In an era of intense political polarization, a critical question is whether these widely-watched "debate" shows foster genuine discussion or have devolved into partisan echo chambers that deepen societal divides. While these programs claim to air competing viewpoints, no large-scale evidence exists to quantify how often hosts and guests actually disagree. Measuring these exchanges is a significant challenge, as live broadcasts contain overlapping speakers, sarcasm, and billions of words of text. To address this gap, we construct the first speaker-resolved map of agreement and disagreement across U.S. cable opinion programming. Our study assembles over 21,000 episodes from 24 flagship shows on Fox News, MSNBC, and CNN from 2010-2024, segmenting them into host-guest turns and labeling 2.13 million turn-pairs using a high-fidelity large-language-model classifier. We present three findings: (1) the proportion of disagreement/debate on prime time shows a consistent downward trend, dropping by roughly one-third between 2017 and 2024; (2) on-air challenge is partisan and asymmetric--conservatives seldom face push-back on Fox, liberals seldom on MSNBC, with CNN declining toward the midpoint; and (3) polarizing issues such as abortion, gun rights, and immigration attract the least disagreement. The work contributes a public corpus, an open-source stance pipeline, and the first longitudinal evidence that televised "debate" is retreating from genuine discussion. By transforming into platforms for partisan affirmation, these shows erode the cross-cutting cleavages essential for a pluralistic society, thereby intensifying affective polarization.

cs.SI↗

arXiv2025

Structural Dynamics of Harmful Content Dissemination on WhatsApp

WhatsApp, a platform with more than two billion global users, plays a crucial role in digital communication, but also serves as a vector for harmful content such as misinformation, hate speech, and political propaganda. This study examines the dynamics of harmful message dissemination in WhatsApp groups, with a focus on their structural characteristics. Using a comprehensive data set of more than 5.1 million messages, including text, images, and videos, collected from approximately 6,000 groups in India, we reconstruct message propagation cascades to analyze dissemination patterns. Our findings reveal that harmful messages consistently achieve greater depth and breadth of dissemination compared to messages without harmful annotations, with videos and images emerging as the primary modes of dissemination. These results suggest a distinctive pattern of dissemination of harmful content. However, our analysis indicates that modality alone cannot fully account for the structural differences in propagation.The findings highlight the critical role of structural characteristics in the spread of these harmful messages, suggesting that strategies targeting structural characteristics of re-sharing could be crucial in managing the dissemination of such content on private messaging platforms.

cs.SI↗

arXiv2025

Toxicity Begets Toxicity: Unraveling Conversational Chains in Political Podcasts

Tackling toxic behavior in digital communication continues to be a pressing concern for both academics and industry professionals. While significant research has explored toxicity on platforms like social networks and discussion boards, podcasts despite their rapid rise in popularity remain relatively understudied in this context. This work seeks to fill that gap by curating a dataset of political podcast transcripts and analyzing them with a focus on conversational structure. Specifically, we investigate how toxicity surfaces and intensifies through sequences of replies within these dialogues, shedding light on the organic patterns by which harmful language can escalate across conversational turns. Warning: Contains potentially abusive/toxic contents.

cs.CL↗

arXiv2025

Analyzing Patterns and Influence of Advertising in Print Newspapers

This paper investigates advertising practices in print newspapers across India using a novel data-driven approach. We develop a pipeline employing image processing and OCR techniques to extract articles and advertisements from digital versions of print newspapers with high accuracy. Applying this methodology to five popular newspapers that span multiple regions and three languages, English, Hindi, and Telugu, we assembled a dataset of more than 12,000 editions containing several hundred thousand advertisements. Collectively, these newspapers reach a readership of over 100 million people. Using this extensive dataset, we conduct a comprehensive analysis to answer key questions about print advertising: who advertises, what they advertise, when they advertise, where they place their ads, and how they advertise. Our findings reveal significant patterns, including the consistent level of print advertising over the past six years despite declining print circulation, the overrepresentation of company ads on prominent pages, and the disproportionate revenue contributed by government ads. Furthermore, we examine whether advertising in a newspaper influences the coverage an advertiser receives. Through regression analyses on coverage volume and sentiment, we find strong evidence supporting this hypothesis for corporate advertisers. The results indicate a clear trend where increased advertising correlates with more favorable and extensive media coverage, a relationship that remains robust over time and across different levels of advertiser popularity.

cs.CY↗

arXiv2025

Global Patterns of Viral Content on WhatsApp

This paper explores the nature and spread of viral WhatsApp content among everyday users in three diverse countries: India, Indonesia, and Colombia. By analyzing hundreds of viral messages collected with participants' consent from private WhatsApp groups, we provide one of the first cross-cultural categorizations of viral content on WhatsApp. Despite the differences in cultural and geographic settings, our findings reveal striking similarities in the types of groups users engage with and the viral content they receive, particularly in the prevalence of misinformation. Our comparative analysis shows that viral content often includes political and religious narratives, with misinformation frequently recirculated despite prior debunking by fact-checking organizations. These parallels suggest that closed messaging platforms like WhatsApp facilitate similar patterns of information dissemination across different cultural contexts. This work contributes to the broader understanding of global digital communication ecosystems and provides a foundation for future research on information flow and moderation strategies in private messaging platforms.

cs.SI↗

arXiv2024

Television Discourse Decoded: Comprehensive Multimodal Analytics at Scale

In this paper, we tackle the complex task of analyzing televised debates, with a focus on a prime time news debate show from India. Previous methods, which often relied solely on text, fall short in capturing the multimodal essence of these debates. To address this gap, we introduce a comprehensive automated toolkit that employs advanced computer vision and speech-to-text techniques for large-scale multimedia analysis. Utilizing state-of-the-art computer vision algorithms and speech-to-text methods, we transcribe, diarize, and analyze thousands of YouTube videos of a prime-time television debate show in India. These debates are a central part of Indian media but have been criticized for compromised journalistic integrity and excessive dramatization. Our toolkit provides concrete metrics to assess bias and incivility, capturing a comprehensive multimedia perspective that includes text, audio utterances, and video frames. Our findings reveal significant biases in topic selection and panelist representation, along with alarming levels of incivility. This work offers a scalable, automated approach for future research in multimedia analysis, with profound implications for the quality of public discourse and democratic debate. To catalyze further research in this area, we also release the code, dataset collected and supplemental pdf.

cs.MM↗

arXiv2024

Examining (Political) Content Consumption on Facebook Through Data Donation

This paper investigates the usage patterns of Facebook among different demographics in the United States, focusing on the consumption of political information and its variability across age, gender, and ethnicity. Employing a novel data donation model, we developed a tool that allows users to voluntarily share their interactions with public Facebook groups and pages, which we subsequently enrich using CrowdTangle. This approach enabled the collection and analysis of a dataset comprising over 1,200 American users. Our findings indicate that political content consumption on Facebook is relatively low, averaging around 17%, and exhibits significant demographic variations. Additionally, we provide insights into the temporal trends of these interactions. The main contributions of this research include a methodological framework for studying social media usage in a privacy-preserving manner, a comprehensive dataset reflective of current engagement patterns, and descriptive insights that highlight demographic disparities and trends over time. This study enhances our understanding of social media's role in information dissemination and its implications for political engagement, offering a valuable resource for researchers and policymakers in a landscape where direct data access is diminishing.

cs.SI↗

arXiv2024

WhatsApp Explorer: A Data Donation Tool To Facilitate Research on WhatsApp

In recent years, reports and anecdotal evidence pointing at the role of WhatsApp in a variety of events, ranging from elections to collective violence, have emerged. While academic research should examine the validity of these claims, obtaining WhatsApp data for research is notably challenging, contrasting with the relative abundance of data from platforms like Facebook and Twitter, where user "information diets" have been extensively studied. This lack of data is particularly problematic since misinformation and hate speech are major concerns in the set of Global South countries in which WhatsApp dominates the market for messaging. To help make research on these questions, and more generally research on WhatsApp, possible, this paper introduces WhatsApp Explorer, a tool designed to enable WhatsApp data collection on a large scale. We discuss protocols for data collection, including potential sampling approaches, and explain why our tool (and adjoining protocol) arguably allow researchers to collect WhatsApp data in an ethical and legal manner, at scale.

cs.SI↗

arXiv2024

Unraveling the Dynamics of Television Debates and Social Media Engagement: Insights from an Indian News Show

The relationship between television shows and social media has become increasingly intertwined in recent years. Social media platforms, particularly Twitter, have emerged as significant sources of public opinion and discourse on topics discussed in television shows. In India, news debates leverage the popularity of social media to promote hashtags and engage users in discussions and debates on a daily basis. This paper focuses on the analysis of one of India's most prominent and widely-watched TV news debate shows: "Arnab Goswami-The Debate". The study examines the content of the show by analyzing the hashtags used to promote it and the social media data corresponding to these hashtags. The findings reveal that the show exhibits a strong bias towards the ruling Bharatiya Janata Party (BJP), with over 60% of the debates featuring either pro-BJP or anti-opposition content. Social media support for the show primarily comes from BJP supporters. Notably, BJP leaders and influencers play a significant role in promoting the show on social media, leveraging their existing networks and resources to artificially trend specific hashtags. Furthermore, the study uncovers a reciprocal flow of information between the TV show and social media. We find evidence that the show's choice of topics is linked to social media posts made by party workers, suggesting a dynamic interplay between traditional media and online platforms. By exploring the complex interaction between television debates and social media support, this study contributes to a deeper understanding of the evolving relationship between these two domains in the digital age. The findings hold implications for media researchers and practitioners, offering insights into the ways in which social media can influence traditional media and vice versa.

cs.SI↗

arXiv2024

Modeling the Impact of Timeline Algorithms on Opinion Dynamics Using Low-rank Updates

Timeline algorithms are key parts of online social networks, but during recent years they have been blamed for increasing polarization and disagreement in our society. Opinion-dynamics models have been used to study a variety of phenomena in online social networks, but an open question remains on how these models can be augmented to take into account the fine-grained impact of user-level timeline algorithms. We make progress on this question by providing a way to model the impact of timeline algorithms on opinion dynamics. Specifically, we show how the popular Friedkin--Johnsen opinion-formation model can be augmented based on aggregate information, extracted from timeline data. We use our model to study the problem of minimizing the polarization and disagreement; we assume that we are allowed to make small changes to the users' timeline compositions by strengthening some topics of discussion and penalizing some others. We present a gradient descent-based algorithm for this problem, and show that under realistic parameter settings, our algorithm computes a $(1+\varepsilon)$-approximate solution in time $\tilde{O}(m\sqrt{n} \lg(1/\varepsilon))$, where $m$ is the number of edges in the graph and $n$ is the number of vertices. We also present an algorithm that provably computes an $\varepsilon$-approximation of our model in near-linear time. We evaluate our method on real-world data and show that it effectively reduces the polarization and disagreement in the network. Finally, we release an anonymized graph dataset with ground-truth opinions and more than 27\,000 nodes (the previously largest publicly available dataset contains less than 550 nodes).

cs.SI↗

arXiv2023

Online Knowledge Production in Polarized Political Memes: The Case of Critical Race Theory

Visual culture has long been deployed by actors across the political spectrum as tools of political mobilization, and have recently incorporated new communication tools, such as memes, GIFs, and emojis. In this study, we analyze the top-circulated Facebook memes relating to critical race theory (CRT) from May 2021 - May 2022 to investigate their visual and textual appeals. Using image clustering techniques and critical discourse analysis, we find that both pro- and anti-CRT memes deploy similar rhetorical tactics to make bifurcating arguments, most of which do not pertain to the academic formulations of CRT. Instead, these memes manipulate definitions of racism and antiracism to appeal to their respective audiences. We argue that labeling such discursive practices as simply a symptom of "post-truth" politics is a potentially unproductive stance. Instead, theorizing the knowledge-building practices of these memes through a lens of political epistemology allows us to understand how they produce meaning.

cs.HC↗