This study conducts a visual analysis of 491 participant‐generated drawings to examine how individuals with varying cybersecurity expertise conceptualize security through visual metaphors and narrative strategies. Using a codebook grounded in visual rhetoric theory, we identified a striking consistency in imagery—particularly the recurring use of locks, shields, and oppositional symbols—across all groups. Findings suggest the existence of shared visual ontologies shaped by both common sense and collective stereotypes. We discuss the implications of these visual conventions for cybersecurity communication and propose future research integrating generative AI tools, audience analysis, and domain knowledge frameworks to uncover underrepresented conceptual gaps.
While there is increased interest in studying the information behavior of individuals engaged in creative pursuits, the body of research on the topic is still small in comparison to the volume of studies concerning the information behavior of scientists, social scientists, and humanists. Two possible and interrelated explanations for this are 1) non‐artists' misunderstanding artists' need for information as a support and inspiration for their work; and 2) the difficulty inherent in studying the creative process, which is highly personal to the artist and often not visible to others. This short paper describes a model for data collection designed to mitigate these challenges; by shifting the preliminary data collection to the artists themselves, diary studies mitigate the challenges presented by attempting to study artists' information behavior in the internal and non‐linear processes of creation. The author designed a study to capture 21 poets' impressions of the role of information in both inspiring and supporting their poetry writing. This paper describes the project design and execution, data collection, and next steps in the author's data analysis process.
As Artificial Intelligence (AI) advancements create new ethical issues and challenge existing legal regulations, traditional ethical principles and data ethics education often fail to reflect these nuances. This study explores how collaborative game design among students with diverse backgrounds fosters transformative learning in data ethics within a college course. It examines a conversation about motivations for enrolling in the course, pre‐course reflections, and post‐course reflections from 32 participants in a semester‐long credit‐based class. Thematic analysis reveals parallels with Mezirow's transformative learning theory (1978) in four key phases of transformative learning. The preliminary findings demonstrate that students not only gained data ethics literacy but also internalized ethical thinking as part of their future goals. This poster highlights the potential of interdisciplinary and collaborative design‐based pedagogy to cultivate digital ethical self‐awareness and ethical judgements when navigating ethical issues in AI or machine learning (ML).
Multimodal disinformation (MD) has become more widespread in social media, raising the importance of information credibility research in this context. Meanwhile, engagement cues like comments and virality metrics may influence online credibility assessment. Using a between‐subjects experiment, we examined the main and interaction effects of multimodality, comments, and virality metrics on disinformation recognition. Our results unexpectedly revealed that MD is easier to recognize than unimodal disinformation. Further, multimodality could attenuate the impact of comments and virality metrics, making people more susceptible to disinformation. Based on these findings, we discussed the theoretical and practical implications.
Unhoused patrons are uniquely vulnerable to marginalizing public library policies. To first capture the scope of the problem and then to offer a lens for productive critique, I perform a content analysis on the behavior policies of all thirty‐five library systems in Los Angeles County, California, USA, describing the methods by which libraries deny unhoused patrons access to library spaces. Then, by applying Iris Marion Young's work on “enabling justice” to these policies, I give a name to these processes: marginalization. I argue that if public libraries are to live up to their name, policies that affect public access need to be viewed through a justice lens. Public libraries that truly are “public” need to design policies that enable justice for all members of their public, and the methods I apply to the Los Angeles County case study offer tools to aid that transformation.
Artificial intelligence (AI) now appears with the search engine results, and this may influence how users interact with search results. This study examined how the incorporation of AI within the information seeking process affected people's health insurance literacy (HIL). Using a HIL measurement, the HIL of young adults was assessed through a simulated Google search for health insurance information. Findings revealed that only 17 of the 41 participants used AI to improve their health insurance literacy, although participants expressed a desire for a simplified explanation of health insurance.
This study explores how the National Central Library in Taiwan leverages a big data service platform to empower public libraries, enabling them to enhance reader participation through data‐driven decisions. The platform systematically processes de‐identified circulation data through comprehensive cleaning, integration, and standardization procedures, enabling libraries to gain actionable insights through advanced visualization tools. This data empowerment initiative reflects the evolution of Taiwan's libraries towards Library 4.0, where analytics capabilities enable libraries to transform from passive information providers to active service innovators, providing personalized, differentiated reader experiences. The platform strengthens public libraries' capacity to promote reading engagement by providing detailed metrics and trend reports, facilitating evidence‐based policy making and service improvements. Through continuous data analysis and visualization, libraries can better understand their communities' reading preferences and adapt their services accordingly. In addition to basic statistical analysis, the platform incorporates and develops data mining techniques to analyze reading behaviors across demographics, offering deeper insights into reader interests and preferences. Furthermore, the study investigates the application of these data mining techniques to facilitate reader resource recommendations, ultimately enhancing library services and promoting a more engaged reading community.
Immersive analytics (IA) leverages immersive and interactive technologies to support data exploration, comprehension, and decision‐making by presenting large volumes of multimodal data in virtual environments. However, the complexity of these systems can pose significant challenges for users, particularly during initial engagement. Effective onboarding is essential to help users acclimate to the environment and learn interaction techniques. In this work, we present findings from a study evaluating three onboarding methods for virtual reality‐based IA systems: a desktop tutorial, in‐Virtual Reality (VR) training, and self‐guided exploration. Results indicate that most participants preferred a combination of all three approaches and emphasized the importance of early access to training components that may require additional time to understand.
Close collaboration between Library and Information Studies (LIS) programs and academic libraries is essential in the age of human‐centered artificial intelligence (AI) to prepare future academic librarians with the knowledge, skills, and mindset needed for modern information services. This panel provides an opportunity for information science educators, researchers, and students to engage with leaders from three U.S. academic libraries. Together, we will explore key challenges currently facing academic libraries and discuss strategies for developing well‐prepared LIS graduates who are equipped to meet these evolving demands. Specifically, this panel will discuss the most pressing challenges facing LIS research and academic libraries in the era of human‐centered AI, important research topics LIS research and academic libraries should prioritize, characteristics of the next generation of academic librarians, and strategies that strengthen collaboration between academic libraries and LIS programs.
As digital infrastructures and methods increasingly shape how cultural memory is preserved, accessed, and interpreted, questions of collaboration and participation have become central to both research and pedagogical practices in digital humanities and cultural heritage contexts. This panel explores the promises and challenges of participatory approaches to digital humanities and cultural heritage work. Bringing together five speakers in various domains of digital humanities, community archives, and digital curation, this panel offers multiple perspectives on how to engage different communities of interest, such as students, interdisciplinary scholars, librarians and practitioners, as well as local communities, in participatory digital humanities and cultural heritage work. Following the individual presentations, panelists will facilitate open discussions with attendees, seeking to collectively explore questions including how to design participatory work in digital humanities and cultural heritage practices, how to engage communities and collaborators in participatory work, and how to address the challenges that emerge in participatory processes. Through this collaborative and interactive approach, this panel seeks to advance knowledge production practices of digital humanities and cultural heritage, advocating for a “participatory future” of digital humanities and cultural heritage work. This panel will be sponsored by ASIS&T SIG‐AVC if accepted.
As machine learning continues to expand across domains, the robust data documentation frameworks and practices have become increasingly important to facilitate transparency. While the development of metadata frameworks is a longstanding focus in the field of information science, little is known about how information science professionals contribute to data documentation framework in machine learning contexts. This study explores the current engagement of information science professionals in developing data documentation frameworks for ML models. We conducted a literature search in a systematic way with the assistance of a generative AI tool and analyzed the authors’ professional backgrounds. Preliminary findings indicate that information science professionals play limited leadership roles in developing data documentation, more often contributing to supporting roles.
Palliative care is frequently misunderstood, yet short videos on social media can help disseminate useful information and build supportive communities. One major challenge is that manually analyzing such content is labor‐intensive and time‐consuming. Meanwhile, large language models (LLMs) show promise for automated content analysis, but their domain‐specific accuracy in this sensitive area remains uncertain. In this study, we propose an iterative LLM‐LLM agentic conversational approach to identify palliative care themes from 56 TikTok videos. We collected video transcripts, metadata, visual labels, and on‐screen text to build a multimodal dataset. Through iterative dialogues between two LLMs, we generated initial themes and refined them via human feedback to address missed dimensions. Our approach identified themes such as Policy, Advocacy, and Access, as well as Emotional Support and Coping while highlighting omissions like Humor and Saying Goodbye, underlining the need for human oversight. Our findings reveal that LLM‐driven automation can reduce annotation workload, but it has limitations in capturing emotional content. The contributions of this work include a new annotated dataset of 242 TikTok videos, a validated LLM‐based thematic analysis pipeline, and evidence that combining automated and human‐in‐the‐loop methods enhances reliability and accuracy in short‐form video analysis.
Disabled and chronically ill populations experience significant barriers to navigating the healthcare system, including communication, attitudinal, and social barriers. Artificial intelligence (AI) may enable disabled and chronically ill individuals to mitigate these barriers. However, most literature exploring the use of AI in healthcare focuses on use by providers and institutions. A growing body of library and information science (LIS) research examines how disabled and chronically ill populations use technologies to manage their health and as tools for information access and communication support (Chen 2016; Costello & Murillo, 2014; Lundy, 2024; St. Jean, 2017). This poster reports developing doctoral student research investigating how disabled and chronically ill populations utilize AI‐enabled chatbots as tools to navigate the healthcare system and manage their health. Semi‐structured interviews are being conducted with a diverse sample of n = 25 disabled and chronically ill participants. Guided by core principles of disability justice, we plan to conduct thematic analysis of the interview data. Our work aims to provide a critical understanding of the chatbot‐facilitated information practices of disabled and chronically ill populations, and to contribute key design considerations for future technologies that support the health and well‐being of these populations.
This study examines whether the semantic meaning of Library of Congress Classification (LCC) class names provides additional insight beyond the system's hierarchical structure for organizing knowledge. Using SBERT, a natural language processing (NLP) model for generating semantic embeddings, we investigate the relationship between the semantic meaning of LCC subclass names and the word usage patterns of the texts assigned to these subclasses. Our results show that although LCC subclass names with similar semantic meanings occasionally correspond to similar word usage patterns, there is no consistent relationship between the two. In contrast, whether two subclasses belong to the same main class reliably predicts the similarity of word usage in the texts assigned to them. While semantic embeddings of LCC subclass names offer intriguing possibilities, our findings indicate that the hierarchical structure of the LCC system remains more robust for knowledge organization.
AI literacy is an emerging research topic, as various AI tools have become embedded in our everyday lives, particularly in the school context. The purposes of this exploratory study are twofold: 1) to investigate high school students’ AI literacy and 2) to explore their experiences using AI tools in the context of coursework. A mixed‐methods approach was employed, combining a survey questionnaire with interviews and information world mapping visual‐elicitation method. The findings indicate that students generally perceived themselves as having a basic understanding of AI and expressed confidence in using AI tools. However, many still relied on human sources, such as peers, to learn how to apply AI in coursework. Notably, gender differences in AI literacy were also observed. These preliminary findings highlight the need for formal AI education and course development in high schools, which may have implications for educators in designing future AI curricula.
Extant research lacks comprehensive identification of distinct AI adopter groups necessary for targeted educational interventions to enhance AI literacy and mitigate the AI divide. Grounded in Social Cognitive Theory (SCT), this study categorizes AI adopters based on social cognitive characteristics, they are, fear of missing out (FoMO), AI attitudes, and self‐efficacy. Further, we profile these groups by AI literacy and educational backgrounds. A survey of 620 participants using K‐means clustering revealed three adopter types: (1) observers, characterized by lower education, AI literacy, FoMO, AI attitudes, and self‐efficacy; (2) seekers, with intermediate educational levels, high FoMO, and strong AI literacy; and (3) professionals, highly educated individuals with low FoMO but high AI literacy. Findings demonstrate how educational disparities shape AI literacy through social cognitive factors. Theoretically, this research introduces an innovative SCT‐based classification of AI adopters, offering practical insights for policymakers and educators to design tailored interventions addressing the AI divide.
Artificial Intelligence (AI) has become prevalent in all sectors of society, including higher education institutions. Many studies have examined iSchools curricula, focusing on areas such as data science, digital humanities, and archival studies. However, few studies have examined AI education at iSchools in the United States (US) and Canada. Research is needed to address AI in information science (IS) education, fueling the conversation about AI across the iSchools' curricula. This study analyzed the AI‐related courses in graduate and undergraduate programs offered by members of the iSchools organization in the United States and Canada. We identified the area(s) and facet(s) covered in each course title and coded them. Of the 51 iSchools, twenty‐nine offered AI‐related courses. The most covered areas include general AI, Machine Learning, Natural Language Processing, Deep Learning, and Robotics. Most courses focus on AI's technical and applied facets, while a few cover the ethical, societal, cultural, and legal facets. Implications include the need for iSchools to offer AI courses that cover aspects beyond the technical, more undergraduate courses, and certificate programs that contribute to educating the labor force that needs upskilling. Drawing from empirical data, this study informs the iSchools' curricula strengths to build on and the gaps to fill and has implications for IS practice.
Japan has the largest ageing population worldwide, with increased numbers of people living with dementia. Dementia legislation aims to disseminate accurate information about dementia, starting with public libraries. However, the policies' effectiveness has not been examined. Therefore, this study analysed how people obtain information about dementia and investigated the significance of disseminating such information in public libraries. An online survey was conducted with 516 people who had cared for someone living with dementia or mild cognitive impairment and who had searched for dementia information at least once in the past year. The respondents often searched online for dementia‐related information, including symptoms and mechanisms. The reliability and accessibility of public libraries were lower than those of hospitals and administrative bodies, but the psychological barriers to obtaining information from public libraries were low. However, some respondents could not find materials about dementia in public libraries, suggesting a need for dementia support services.
This paper analyzes 25 health information behavior studies published in the period of 15 years (2009‐2023) using eye‐tracking as the research methodology. Eye‐tracking technology has become a valuable tool to study information behavior. This study examines the ways of adopting eye trackers in health information behavior research and how the research is designed with an extended method the technology provides. This paper contributes to understanding methodological trends for health information behavior research.
Palliative care is a health service to patients and caregivers to improve quality of life when facing a terminal illness. In this work‐in‐progress paper, we analyzed secondary data, namely 46 TikToks thematizing palliative care, to gain initial insights into the creators and types of content in recent years. We developed a codebook based on inductive and deductive methods to conduct content analysis. Twenty‐four TikToks were posted by healthcare professionals, while thirteen were created by patients. Six TikToks were created by caregivers and three TikToks were posted by news agencies. While content of TikToks varied, around 70% of 46 TikToks were categorized as Personal Experience shared by healthcare professionals, patients and caregivers. Additionally, our goal was to explore research potential concerning affective information behavior and the development of para‐social and/or cyber social interactions and relationships triggered by highly vulnerable and/or emotionally charged content. Therefore, we manually collected and analyzed 20 user comments for each of the five TikToks selected out of the 46 TikToks. These TikTok users provided emotional support by expressing sympathy and offering prayers; they also asked questions.
