UMSI at ASIS&T 2026
Thursday, 11/05/2026
By Noor HindiThe 89th Annual Meeting of the Association for Information Science and Technology (ASIS&T) will be held in Bangkok, Thailand from November 6-10. Several University of Michigan School of Information researchers will be presenting their work.
Workshop: Individual Behaviour to Ecosystems: Research Agendas Spanning Scales, Methods, Disciplines
Libby Hemphill, Alicia Wanless, Megan Threats, Amelia Acker
The information ecosystems people inhabit shape what they seek, consume, and believe—yet researchers who study individual information behaviour and those who study national-scale media environments rarely work together. This workshop brings those researchers together to outline research agendas about information ecosystems that span scales, methods, and scholarly communities.
Researchers often study a single place people access information (e.g., social media), a single level of an information ecosystem (e.g., community), or a single type of information sought (e.g., health). This workshop aims to move beyond isolated studies of information behaviour by fostering a systemic, holistic conversation. We invite scholars to examine all levels of the information ecosystem.
We welcome researchers who wish to understand and articulate how individual behaviours and environmental factors influence one another. They may be studying information seeking or retrieval, algorithmic curation, narrative diffusion, media literacy, and the spread of disinformation using computational, behavioural, or humanistic methods.
Our primary goal is to establish a collaborative research agenda. Participants will work together to define at least one concrete research project and team. We will draft a white paper based on the workshop and include participants. The workshop will be held in-person.
Precarious Information Practices: Disabled and Chronically Ill Communities and the Use of AI-Enabled Chatbots to Navigate Epistemic Injustice in Healthcare Systems
Artificial intelligence-enabled chatbots are transforming how individuals access and interact with health information. Little is known about how these technologies are used by individuals navigating barriers within healthcare systems. This study examines how disabled and chronically ill individuals use chatbots to navigate healthcare systems that marginalize their embodied knowledge. Semi-structured interviews were conducted with 25 disabled and/or chronically ill participants and analyzed using thematic analysis informed by disability justice principles. Findings show participants used chatbots as a sociotechnical mechanism to bridge informational and communication divides, enabling participation in a system that marginalized their ways of knowing. In response to ableism and fluctuating health states, participants used chatbots to develop knowledge, translate embodied experiences into clinically legible language, and engage in self-advocacy during clinical encounters. We define these as precarious information practices – agentic, yet unstable practices enacted to survive under conditions of information marginalization. Our findings highlight the dual role of chatbots as both enabling and precarious health information technologies among disabled and chronically ill communities.
Tinkerers, Techies, and Skeptics: Fixing, Breaking, Making, Reconfiguring, and Building Informal Infrastructures
Eric Meyer, Dang Nguyen, Michael Twidale, Morgan Wofford
Socio-technical platforms and tools have become more locked-down, tightly controlled, and one-size-fits-all in recent years, which appeals to people who want to open an app or a device and have it ‘just work’. Others want to open the black box, see what is inside, shift things around, and discover new, creative uses, a.k.a. tinkering.
In this panel, four complementary views on tinkering are shared in the interests of clarifying the role of tinkering in how people engage with information and other socio-technical constructs. Meyer will explain his model of tinkering through a case from home automation. Nguyen examines Vietnamese phone farming as a case of how socio-technical systems are reconfigured to generate new value beyond their intended design. Twidale explores ‘techies’ vs. ‘non-techies’, and how we can learn from how techies approach challenges to make the experience of non-techies more satisfying. Wofford examines anthropogenic climate crisis skeptics who reuse open data, and, in doing so, construct an alternative knowledge infrastructure that repurposes data beyond its original design.
The panel will invite the audience to discuss fixing, breaking, making, reconfiguring, and building informal infrastructures and to consider how these ideas can inform information science more widely.
Pronouns Through the Lens of Queer Information Practices
Isabelle “Beck” Mallwitz, Patricia Garcia
Pronouns have become a contested site in broader debates over trans rights, yet little information science research has examined how people make sense of and use pronoun-related information in social contexts. This paper introduces queer information as a theoretical concept that describes relational, identity-based information created for, by, and about queer people, and uses pronouns as an empirical case to explore it. Through 12 semi-structured interviews with cis, trans, and gender-nonconforming individuals, we examine how people conceptualize gendered pronouns, how those conceptualizations shape their uses and attitudes, and how pronoun circles function as a public queer information practice. We conclude by discussing the contextual, relational, and imaginative qualities of queer information and its implications for understanding how information practices shape core sociopolitical issues such as identity and community membership.
Using Data Journeys to Understand the Impact of Changing Infrastructural Contexts on Data Reuse
Morgan F. Wofford, Nicholas Hoffman, Jessica Blois, Andrea K. Thomer
Data reuse entails moving data from one context into another. This recontextualization can result in the loss of important information, which can lead to unexpected consequences, including harm. Methods are needed to better analyze the impact of changing data contexts over time. We take a data journeys approach to illustrate data recontextualization across two cases: the reuse of climate science data by anthropogenic climate crisis skeptics, and the reuse of archaeological data by paleoecologists. By doing so, we set the stage to better understand how infrastructures shape the data they provide and archive.
Beyond Dataset-Level Description: Automated Keyword Generation for Variable Discovery and Harmonization in Social Science Data
Metadata are foundational to data discovery and reuse, yet variable-level documentation remains prohibitively labor-intensive at scale. We experimentally evaluated large language model (LLM) approaches to automatically generating keyword metadata for social science survey variables using a 2×2 design: zero-shot versus fine-tuned generation, and open versus controlled vocabulary output. Using variables from the Social Science Variables Database (SSVD), we assessed outputs on language quality metrics and a variable matching task. Zero-shot open-vocabulary generation achieved the strongest overall variable matching performance and likely supports free-text search and findability. Fine-tuned controlled-vocabulary generation best discriminated highly similar variables and supports interoperability and harmonization. These conditions offer complementary FAIR affordances from a single metadata generation pipeline. Practitioners should select generation strategies based on downstream application, and LLM-generated metadata can meaningfully improve variable-level findability and interoperability in large social science data archives.
RELATED
Check out more UMSI research by subscribing to our free research roundup newsletter!