arXiv subjects

Ho-Chun Herbert Chang

Publications and source records attributed to Ho-Chun Herbert Chang.

At least 19 recordsLinked to original sources

arXiv2026

The 2026 World Cup, Anime Edition: Production, Engagement, and Circulation of Generative Memetic Sports Content

Generative AI gives audiences new ways to portray real sporting figures through familiar fictional worlds. Expanded opportunities to create, however, need not broaden cultural representation or distribute attention evenly. We examine this tension through the supply, demand, and circulation of anime style football videos on Instagram. Using 2,021 posts and associated video and account data, we combine caption entity extraction, human-reviewed visual identity clustering, engagement analysis, and content-reuse networks. Representation is uneven across athletes, while Blue Lock, Naruto, and One Piece account for 60.1% of explicit franchise mentions. Among posts referencing generative tools, rankings by total engagement differ from rankings by average engagement per video, distinguishing aggregate visibility from response to individual posts. Reuse is extensive: 772 of 1,548 evaluated videos belong to 124 detected groups, including 71 spanning multiple known accounts. Shared material connects a large account component alongside smaller groups, and later observed uploads receive lower median normalized engagement than first observed appearances. These findings characterize generative cultural crossover as a practice organized around concentrated cultural resources, uneven attention, and extensive recirculation. More broadly, they distinguish opportunities for visual authorship from the processes through which particular representations become widely visible.

cs.SI↗

arXiv2026

Who Anchors AI Overviews in Health? Baidu, Google, and the Geography of Authority

Artificial intelligence is being rapidly incorporated into traditional search systems, yet scant work audits the information disparities across platforms, geography, and languages. We address this gap by comparing Google and Baidu's AI Overview systems for health queries, and measure informational anchors that emerge. Auditing 1,920 health queries across 12 countries and 4 languages, we find that Google and Baidu exhibit vertical integration, routing users toward their own company platforms in AI Overviews rather than a diverse set of primary sources. Smaller, lower-localization countries receive fewer domestically sourced references for health queries. Issuing the same query in a country's official language rather than English raises the share of locally sourced citations approximately 3.5- to 13.5-fold. Comparing queries across health topics of varying severity and controversy, including Traditional Chinese Medicine as an example, we also show that health disclaimers are multidimensional and vary across language and culture. We discuss how generative search influences access to health information, and the urgent need for culturally-aware oversight of these systems that influence critical health decisions.

cs.IR↗

arXiv2026

Modest Algorithmic Mediation can Maximize Topical Diversity in Hybrid Human-AI Systems

In the artificial intelligence (AI) era, the rise of algorithmic feeds has fundamentally transformed information diffusion on social media. While early platforms organized visibility through explicit social networks, contemporary systems mediate exposure through intelligent recommender algorithms that personalize attention. This paper examines how the social network and algorithmic architecture jointly shape the diversity of information sharing. Analysis of 18,076 users active throughout 2014--2018 shows that the topical diversity of sharing rose and then plateaued after the introduction of algorithmic ranking in 2016 while its inequality across users emerged alongside it. To this end, we introduce a hybrid human-AI information diffusion model in which information exposure is governed by a parameterized mixture of social propagation through the user-following network and algorithmic recommendation. Both qualitative analysis and simulations show that the effect of algorithmic mediation is non-monotonic. Modest mediation can raise average diversity and reduce inequality relative to a purely network-driven baseline, whereas strong mediation reduces diversity and concentrates it among fewer users. Fitting the model to four years of data yields a mediation share that increases from zero before 2016 to approximately 0.50 by 2018, a level that exceeds the compensation point of equality while remaining within the diversity-enhancing range. These results identify the conditions under which recommendation broadens rather than narrows exposure and provide a unified framework for information diffusion in hybrid human-AI systems.

cs.SI↗

arXiv2026

The Meme Is the Message: Generative Memesis and AI Visuals in the 2024 USA Presidential Elections

Visual content on social media has become increasingly influential in shaping political discourse and civic engagement, but it also limits participation due to the increased cost of multimedia production. In tandem, the growth of generative AI provides novel ways for citizens to participate in politics by lowering these costs. Drawing on a dataset of 239,526 Instagram images, we analyze the effects of synthetic images during the 2024 United States presidential election, using a multimodal workflow combining computer vision, large language models, and facial affect analysis. Results show that meme format is a stronger predictor of engagement than AI-generated content alone. However, AI-generated memes yield a significant interaction effect, suggesting synergistic increases in engagement when synthetic imagery is integrated with memes through human curation. We also characterize how users curate images. Partisans use AI in different ways: Democrat-leaning users tend to use it for in-group support, whereas Republican-leaning users more often employ it for out-group attacks. Users generally select happier synthetic faces compared to real photographs. We define generative memesis as a mode of communication in which memes are no longer shared person-to-person, but mediated by AI through customized visuals. We discuss how generative AI may empower civic participation, the bifurcation of content production and curation, and its implications for in the history of novel technologies and participatory culture.

cs.CY↗

arXiv2026

The First Mass Protest on Threads: Multimodal Mobilization and AI-Generated Visuals in Taiwan's Bluebird Movement

The 2024 Bluebird Movement in Taiwan marked one of the largest youth-led protests in the country's democratic history, mobilizing over 100,000 demonstrators in response to parliamentary reforms. Unlike the 2014 Sunflower Movement, Bluebird unfolded within a transformed digital environment dominated by Threads, Meta's new microblogging platform that uniquely draws 24% of its global traffic from Taiwan. Leveraging a dataset of 62,321 posts and 21,572 images, this study analyzes how protest communication developed across textual and visual modalities. We combine LLM zero-shot annotation, gradient-boosting trees, and SHAP explainers to disambiguate the supply and demand of attention. Results reveal three dynamics: (1) partisan asymmetries between algorithmic exposure and user endorsement, with anti-DPP content surfaced more widely but anti-KMT and pro-DPP content more actively recirculated; (2) textual repertoires centered on commemorations, personal testimonies, and calls to action as key drivers of virality; and (3) a bifurcation in visual strategies, where human photographs concentrated exposure and discussion, while AI-generated animal and plant symbols circulated as mobilization tools and partisan attacks. These findings demonstrate how Threads functioned as both an amplifier and filter of democratic contention, extending theories of emotional and visual contagion by showing how generative AI reshapes symbolic repertoires in contemporary protest through what we term kawaii toxicity, political attacks cloaked in aesthetics of cuteness.

cs.CY↗

arXiv2026

Interpreting Multimodal Communication at Scale in Short-Form Video: Visual, Audio, and Textual Mental Health Discourse on TikTok

Short-form video platforms integrate text, visuals, and audio into complex communicative acts, yet existing research analyzes these modalities in isolation, lacking scalable frameworks to interpret their joint contributions. This study introduces a pipeline combining automated multimodal feature extraction with Shapley value-based interpretability to analyze how text, visuals, and audio jointly influence engagement. Applying this framework to 162,965 TikTok videos and 814,825 images about social anxiety disorder (SAD), we find that facial expressions outperform textual sentiment in predicting viewership, informational content drives more attention than emotional support, and cross-modal synergies exhibit threshold-dependent effects. These findings demonstrate how multimodal analysis reveals interaction patterns invisible to single-modality approaches. Methodologically, we contribute a reproducible framework for interpretable multimodal research applicable across domains; substantively, we advance understanding of mental health communication in algorithmically mediated environments.

cs.MM↗

arXiv2025

Food as Soft Power: Taiwanese Gastrodiplomacy on Social Media and Algorithmic Suppression

Social media platforms have become pivotal for projecting national identity and soft power in an increasingly digital world. This study examines the digital manifestation of Taiwanese gastrodiplomacy by focusing on bubble tea -- a culturally iconic beverage -- leveraging a dataset comprising 107,169 posts from the popular lifestyle social media platform Instagram. Including 315,279,227 engagements, 4,756,320 comments, and 8,097,260,651 views over five full years (2020-2024), we investigate how social media facilitates discussion about Taiwanese cuisine and contributes to Taiwan's digital soft power. Our analysis reveals that bubble tea consistently emerges as the dominant representation of Taiwanese cuisine across Meta's Instagram channels. However, this dominance also indicates vulnerability in gastrodiplomatic strategy compared to other countries. Additionally, we find evidence that Instagram suppresses bubble tea posts mentioning Taiwan by 1,200% -- roughly a twelve-fold decrease in exposure -- relative to posts without such mentions. Crucially, we observe a significant drop in the number of posts, views, and engagement following Lai's inauguration in May 2024. This study ultimately contributes to understanding how digital platforms can enable or disable gastrodiplomacy, soft power, and cultural diplomacy while highlighting the need for greater algorithmic transparency. By noting Taiwan's bubble tea's digital engagement and footprint, critical insights are brought for nations seeking to leverage soft power through gastronomic means in a politicized digital era and researchers trying to better understand algorithmic suppression.

cs.SI↗

arXiv2025

Measuring Social Influence with Networked Synthetic Control

Measuring social influence is difficult due to the lack of counter-factuals and comparisons. By combining machine learning-based modeling and network science, we present general properties of social value, a recent measure for social influence using synthetic control applicable to political behavior. Social value diverges from centrality measures on in that it relies on an external regressor to predict an output variable of interest, generates a synthetic measure of influence, then distributes individual contribution based on a social network. Through theoretical derivations, we show the properties of SV under linear regression with and without interaction, across lattice networks, power-law networks, and random graphs. A reduction in computation can be achieved for any ensemble model. Through simulation, we find that the generalized friendship paradox holds -- that in certain situations, your friends have on average more influence than you do.

cs.SI↗

arXiv2025

The Niche Connectivity Paradox: Multichrome Contagions Overcome Vaccine Hesitancy more effectively than Monochromacy

The rise of vaccine hesitancy has caused a resurgence of vaccine-preventable diseases such as measles and pertussis, alongside widespread skepticism and refusals of COVID-19 vaccinations. While categorizing individuals as either supportive of or opposed to vaccines provides a convenient dichotomy of vaccine attitudes, vaccine hesitancy is far more complex and dynamic. It involves wavering individuals whose attitudes fluctuate -- those who may exhibit pro-vaccine attitudes at one time and anti-vaccine attitudes at another. Here, we identify and analyze multichrome contagions as potential targets for intervention by leveraging a dataset of known pro-vax and anti-vax Twitter users ($n =135$ million) and a large COVID-19 Twitter dataset ($n = 3.5$ billion; including close analysis of $1,563,472$ unique individuals). We reconstruct an evolving multiplex sentiment landscape using top co-spreading issues, characterizing them as monochrome and multichrome contagions, based on their conceptual overlap with vaccination. We demonstrate switchers as deliberative: they are more moderate, engage with a wider range of topics, and occupy more central positions in their networks. Further examination of their information consumption shows that their discourse often engages with progressive issues such as climate change, which can serve as avenues for multichrome contagion interventions to promote pro-vaccine attitudes. Using data-driven intervention simulations, we demonstrate a paradox of niche connectivity, where multichrome contagions with fragmented, non-overlapping communities generate the highest levels of diffusion for pro-vaccine attitudes. Our work offers insights into harnessing synergistic hitchhiking effect of multichrome contagions to drive desired attitude and behavior changes in network-based interventions, particularly for overcoming vaccine hesitancy.

cs.SI↗

arXiv2025

Faces Speak Louder Than Words: Emotions Versus Textual Sentiment in the 2024 USA Presidential Election

Sentiment analysis of textual content has become a well-established solution for analyzing social media data. However, with the rise of images and videos as primary modes of expression, more information on social media is conveyed visually. Among these, facial expressions serve as one of the most direct indicators of emotional content in images. This study analyzes a dataset of Instagram posts related to the 2024 U.S. presidential election, spanning April 5, 2024, to August 9, 2024, to compare the relationship between textual and facial sentiment. Our findings reveal that facial expressions align with text sentiment, where positive sentiment aligns with happiness, although neutral and negative facial expressions provide critical information beyond negative valence. Furthermore, during politically significant events such as Donald Trump's conviction and assassination attempt, posts depicting Trump showed a 12% increase in negative sentiment. Crucially, Democrats use their opponent's fear to depict weakness, whereas Republicans use their candidate's anger to depict resilience. Our research highlights the potential of integrating facial expression analysis with textual sentiment analysis to uncover deeper insights into social media dynamics.

cs.SI↗

arXiv2024

Gender Inequalities in Content Collaborations: Asymmetric Creator Synergy and Symmetric Audience Biases

Content-creator collaborations are a widespread strategy for enhancing digital viewership and revenue. While existing research has explored the efficacy of collaborations, few have looked at inequities in collaborations, particularly from the perspective of the supply and demand of attention. Leveraging 42,376 videos and 6,117,441 comments from YouTube (across 150 channels and 3 games), this study examines gender inequality in collaborative environments. Utilizing Shapley value, a tool from cooperative game theory, results reveal dominant in-group collaborations based on in-game affordances. However, audience responses are aligned across games, reflecting symmetric biases across the gaming communities, with comments focusing more on peripherals than actual gameplay for women. We find supply-side asymmetries exist along with demand-side symmetries. Our results engage with the larger literature on digital and online biases, highlighting how genre and affordances moderate gendered collaboration, the direction of inequality, and contributing a general framework to quantify synergy across collaborations.

cs.CY↗

arXiv2024

WinoQueer: A Community-in-the-Loop Benchmark for Anti-LGBTQ+ Bias in Large Language Models

We present WinoQueer: a benchmark specifically designed to measure whether large language models (LLMs) encode biases that are harmful to the LGBTQ+ community. The benchmark is community-sourced, via application of a novel method that generates a bias benchmark from a community survey. We apply our benchmark to several popular LLMs and find that off-the-shelf models generally do exhibit considerable anti-queer bias. Finally, we show that LLM bias against a marginalized community can be somewhat mitigated by finetuning on data written about or by members of that community, and that social media text written by community members is more effective than news text written about the community by non-members. Our method for community-in-the-loop benchmark development provides a blueprint for future researchers to develop community-driven, harms-grounded LLM benchmarks for other marginalized communities. Note: This version corrects a bug found in evaluation code after publication. General findings have not changed, but tables 5 and 6 and figure 1 have been corrected.

cs.CL↗

arXiv2024

LLMs with Personalities in Multi-issue Negotiation Games

Powered by large language models (LLMs), AI agents have become capable of many human tasks. Using the most canonical definitions of the Big Five personality, we measure the ability of LLMs to negotiate within a game-theoretical framework, as well as methodological challenges to measuring notions of fairness and risk. Simulations (n=1,500) for both single-issue and multi-issue negotiation reveal increase in domain complexity with asymmetric issue valuations improve agreement rates but decrease surplus from aggressive negotiation. Through gradient-boosted regression and Shapley explainers, we find high openness, conscientiousness, and neuroticism are associated with fair tendencies; low agreeableness and low openness are associated with rational tendencies. Low conscientiousness is associated with high toxicity. These results indicate that LLMs may have built-in guardrails that default to fair behavior, but can be "jail broken" to exploit agreeable opponents. We also offer pragmatic insight in how negotiation bots can be designed, and a framework of assessing negotiation behavior based on game theory and computational social science.

cs.CL↗

arXiv2023

What are Your Pronouns? Examining Gender Pronoun Usage on Twitter

Stating your gender pronouns, along with your name, is becoming the new norm of self-introductions at school, at the workplace, and online. The increasing prevalence and awareness of nonconforming gender identities put discussions of developing gender-inclusive language at the forefront. This work presents the first empirical research on gender pronoun usage on large-scale social media. Leveraging a Twitter dataset of over 2 billion tweets collected continuously over two years, we find that the public declaration of gender pronouns is on the rise, with most people declaring as using she series pronouns, followed by he series pronouns, and a smaller but considerable amount of non-binary pronouns. From analyzing Twitter posts and sharing activities, we can discern users who use gender pronouns from those who do not and also distinguish users of various gender identities. We further illustrate the relationship between explicit forms of social network exposure to gender pronouns and their eventual gender pronoun adoption. This work carries crucial implications for gender-identity studies and initiates new research directions in gender-related fairness and inclusion, as well as support against online harassment and discrimination on social media.

cs.SI↗

arXiv2023

Identity Collapse? Realignment of Taiwanese Voters in the 2024 Presidential Elections on Social Media

The 2024 Taiwanese Presidential Election is not just a critical geopolitical event, it also engages with long-standing debate in politics regarding the factors that lead to the rise of new political parties and candidates. In 2021, the Economist called Taiwan "the most dangerous place on earth" due to its critical role in a fragile supply chain. Additionally, a four-candidate race has emerged in a traditionally bipartisan election which begs the question: how will voters realign given the choice of four candidates? Leveraging more than a million posts on social media, we analyze user (predominantly Taiwanese) discourse and engagement along the axes of national identity, issue topic, and partisan alignment. Results reveal alternative candidates (Ko and Gou) draw attention from the fringes rather than the center relative to national identity, and traditional candidates derive more engagement from the traditional media and salience to geopolitical issues. Crucially, in-group references generate more engagement than out-group references, contrary to Western-based studies. We discuss how the dissolution of Taiwan's single-issue society may not just lead to more viable candidates and multi-issue discourse, but the misalignment of national and partisan identity may heal deep-seated partisan cleavages.

cs.CY↗

arXiv2023

Parasocial diffusion: K-pop fandoms help drive COVID-19 public health messaging on social media

We examine an unexpected but significant source of positive public health messaging during the COVID-19 pandemic -- K-pop fandoms. Leveraging more than 7 million tweets related to mask-wearing and K-pop between March 2020 and December 2021, we analyzed the online spread of the hashtag \#WearAMask and vaccine-related tweets amid anti-mask sentiments and public health misinformation. Analyses reveal the South Korean boyband BTS as one of the most significant driver of health discourse. Tweets from health agencies and prominent figures that mentioned K-pop generate 111 times more online responses compared to tweets that did not. These tweets also elicited strong responses from South America, Southeast Asia, and rural States -- areas often neglected in Twitter-based messaging by mainstream social media campaigns. Network and temporal analysis show increased use from right-leaning elites over time. Mechanistically, strong-levels of parasocial engagement and connectedness allow sustained activism in the community. Our results suggest that public health institutions may leverage pre-existing audience markets to synergistically diffuse and target under-served communities both domestically and globally, especially during health crises such as COVID-19.

cs.SI↗

arXiv2023

Complex Systems of Secrecy: The Offshore Networks of Oligarchs

Following the invasion of Ukraine, the US, UK, and EU governments--among others--sanctioned oligarchs close to Putin. This approach has come under scrutiny, as evidence has emerged of the oligarchs' successful evasion of these punishments. To address this problem, we analyze the role of an overlooked but highly influential group: the secretive professional intermediaries who create and administer the oligarchs' offshore financial empires. Drawing on the Offshore Leaks Database provided by the International Consortium of Investigative Journalists (ICIJ), we examine the ties linking offshore expert advisors (lawyers, accountants, and other wealth management professionals) to ultra-high-net-worth individuals from four countries: Russia, China, the United States, and Hong Kong. We find that resulting nation-level "oligarch networks" share a scale-free structure characterized by heterogeneity of heavy-tailed degree distributions of wealth managers; however, network topologies diverge across clients from democratic versus autocratic regimes. While generally robust, scale-free networks are fragile when targeted by attacks on highly-connected nodes. Our "knock-out" experiments pinpoint this vulnerability to the small group of wealth managers themselves, suggesting that sanctioning these professional intermediaries may be more effective and efficient in disrupting dark finance flows than sanctions on their wealthy clients. This vulnerability is especially pronounced amongst Russian oligarchs, who concentrate their offshore business in a handful of boutique wealth management firms. The distinctive patterns we identify suggest a new approach to sanctions, focused on expert intermediaries to disrupt the finances and alliances of their wealthy clients. More generally, our research contributes to the larger body of work on complexity science and the structures of secrecy.

cs.SI↗

arXiv2022

Towards WinoQueer: Developing a Benchmark for Anti-Queer Bias in Large Language Models

This paper presents exploratory work on whether and to what extent biases against queer and trans people are encoded in large language models (LLMs) such as BERT. We also propose a method for reducing these biases in downstream tasks: finetuning the models on data written by and/or about queer people. To measure anti-queer bias, we introduce a new benchmark dataset, WinoQueer, modeled after other bias-detection benchmarks but addressing homophobic and transphobic biases. We found that BERT shows significant homophobic bias, but this bias can be mostly mitigated by finetuning BERT on a natural language corpus written by members of the LGBTQ+ community.

cs.CL↗