Kaiping Chen

Kaiping Chen, Anqi Shao, Jirayu Burapacheep et al.

Autoregressive language models, which use deep learning to produce human-like texts, have become increasingly widespread. Such models are powering popular virtual assistants in areas like smart health, finance, and autonomous driving. While the parameters of these large language models are improving, concerns persist that these models might not work equally for all subgroups in society. Despite growing discussions of AI fairness across disciplines, there lacks systemic metrics to assess what equity means in dialogue systems and how to engage different populations in the assessment loop. Grounded in theories of deliberative democracy and science and technology studies, this paper proposes an analytical framework for unpacking the meaning of equity in human-AI dialogues. Using this framework, we conducted an auditing study to examine how GPT-3 responded to different sub-populations on crucial science and social topics: climate change and the Black Lives Matter (BLM) movement. Our corpus consists of over 20,000 rounds of dialogues between GPT-3 and 3290 individuals who vary in gender, race and ethnicity, education level, English as a first language, and opinions toward the issues. We found a substantively worse user experience with GPT-3 among the opinion and the education minority subpopulations; however, these two groups achieved the largest knowledge gain, changing attitudes toward supporting BLM and climate change efforts after the chat. We traced these user experience divides to conversational differences and found that GPT-3 used more negative expressions when it responded to the education and opinion minority groups, compared to its responses to the majority groups. We discuss the implications of our findings for a deliberative conversational AI system that centralizes diversity, equity, and inclusion.

Isabel Villanueva, Tara Bobinac, Binwei Yao et al.

Despite increasing AI chatbot deployment in public discourse, empirical evidence on their capacity to foster intercultural empathy remains limited. Through a randomized experiment, we assessed how different AI deliberation approaches--cross-cultural deliberation (presenting other-culture perspectives), own-culture deliberation (representing participants' own culture), and non-deliberative control--affect intercultural empathy across American and Latin American participants. Cross-cultural deliberation increased intercultural empathy among American participants through positive emotional engagement, but produced no such effects for Latin American participants, who perceived AI responses as culturally inauthentic despite explicit prompting to represent their cultural perspectives. Our analysis of participant-driven feedback, where users directly flagged and explained culturally inappropriate AI responses, revealed systematic gaps in AI's representation of Latin American contexts that persist despite sophisticated prompt engineering. These findings demonstrate that current approaches to AI cultural alignment--including linguistic adaptation and explicit cultural prompting--cannot fully address deeper representational asymmetries in AI systems. Our work advances both deliberation theory and AI alignment research by revealing how the same AI system can simultaneously promote intercultural understanding for one cultural group while failing for another, with critical implications for designing equitable AI systems for cross-cultural democratic discourse.

Sha Luo, Sang Jung Kim, Zening Duan et al.

Refusal behavior by Large Language Models is increasingly visible in content moderation, yet little is known about how refusals vary by the identity of the user making the request. This study investigates refusal as a sociotechnical outcome through a counterfactual persona design that varies gender identity--including male, female, non-binary, and transgender personas--while keeping the classification task and visual input constant. Focusing on a vision-language model (GPT-4V), we examine how identity-based language cues influence refusal in binary gender classification tasks. We find that transgender and non-binary personas experience significantly higher refusal rates, even in non-harmful contexts. Our findings also provide methodological implications for equity audits and content analysis using LLMs. Our findings underscore the importance of modeling identity-driven disparities and caution against uncritical use of AI systems for content coding. This study advances algorithmic fairness by reframing refusal as a communicative act that may unevenly regulate epistemic access and participation.

Zening Duan, Anqi Shao, Yicheng Hu et al.

While researchers often study message features like moral content in text, such as party manifestos and social media, their quantification remains a challenge. Conventional human coding struggles with scalability and intercoder reliability. While dictionary-based methods are cost-effective and computationally efficient, they often lack contextual sensitivity and are limited by the vocabularies developed for the original applications. In this paper, we present an approach to construct vec-tionary measurement tools that boost validated dictionaries with word embeddings through nonlinear optimization. By harnessing semantic relationships encoded by embeddings, vec-tionaries improve the measurement of message features from text, especially those in short format, by expanding the applicability of original vocabularies to other contexts. Importantly, a vec-tionary can produce additional metrics to capture the valence and ambivalence of a message feature beyond its strength in texts. Using moral content in tweets as a case study, we illustrate the steps to construct the moral foundations vec-tionary, showcasing its ability to process texts missed by conventional dictionaries and word embedding methods and to produce measurements better aligned with crowdsourced human assessments. Furthermore, additional metrics from the vec-tionary unveiled unique insights that facilitated predicting outcomes such as message retransmission.

Kaiping Chen, Yepeng Jin, Anqi Shao

Misinformation and intergroup bias are two pathologies challenging informed citizenship. This paper examines how identity language is used in misinformation and debunking messages about controversial science on Chinese digital public sphere, and their impact on how the public engage with science. We collected an eight-year time series dataset of public discussion (N=6039) on one of the most controversial science issues in China (GMO) from a popular Q&A platform, Zhihu. We found that both misinformation and debunking messages use a substantial amount of group identity languages when discussing the controversial science issue, which we define as science factionalism -- discussion about science is divided by factions that are formed upon science attitudes. We found that posts that use science factionalism receive more digital votes and comments, even among the science-savvy community in China. Science factionalism also increases the use of negativity in public discourse. We discussed the implications of how science factionalism interacts with the digital attention economy to affect public engagement with science misinformation.

Kaiping Chen, Sang Jung Kim, Qiantong Gao et al.

Recent years have witnessed an explosion of science conspiracy videos on the Internet, challenging science epistemology and public understanding of science. Scholars have started to examine the persuasion techniques used in conspiracy messages such as uncertainty and fear yet, little is understood about the visual narratives, especially how visual narratives differ in videos that debunk conspiracies versus those that propagate conspiracies. This paper addresses this gap in understanding visual framing in conspiracy videos through analyzing millions of frames from conspiracy and counter-conspiracy YouTube videos using computational methods. We found that conspiracy videos tended to use lower color variance and brightness, especially in thumbnails and earlier parts of the videos. This paper also demonstrates how researchers can integrate textual and visual features in machine learning models to study conspiracies on social media and discusses the implications of computational modeling for scholars interested in studying visual manipulation in the digital era. The analysis of visual and textual features presented in this paper could be useful for future studies focused on designing systems to identify conspiracy content on the Internet.

Kaiping Chen, Anfan Chen, Jingwen Zhang et al.

This paper studies conspiracy and debunking narratives about COVID-19 origination on a major Chinese social media platform, Weibo, from January to April 2020. Popular conspiracies about COVID-19 on Weibo, including that the virus is human-synthesized or a bioweapon, differ substantially from those in the US. They attribute more responsibility to the US than to China, especially following Sino-US confrontations. Compared to conspiracy posts, debunking posts are associated with lower user participation but higher mobilization. Debunking narratives can be more engaging when they come from women and influencers and cite scientists. Our findings suggest that conspiracy narratives can carry highly cultural and political orientations. Correction efforts should consider political motives and identify important stakeholders to reconstruct international dialogues toward intercultural understanding.