Co-Data
Cultivating Effective Human-LLM Collaboration for Collaborative Data Processing
Amedeo Pachera (Université Claude Bernard Lyon 1)
Andrea Mauri (Université Claude Bernard Lyon 1)
Kashif Imteyaz (Northeastern University)
Jie Yang (TU Delft - Electrical Engineering, Mathematics and Computer Science)
Eric Umuhoza (Carnegie Mellon University Africa)
Angela Bonifati (Université Claude Bernard Lyon 1, Institut Universitaire de France)
Michal Lahav (Google)
Nitesh Goyal (Google)
More Info
expand_more
Other than for strictly personal use, it is not permitted to download, forward or distribute the text or part of it, without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license such as Creative Commons.
Abstract
Data work is increasingly collaborative and multidisciplinary, yet teams struggle with mismatched semantics, uneven data literacy, and variable trust in automation. Large Language Models (LLMs) now assist with cleaning, integration, annotation, and querying, but their role in mediating collaboration, aligning goals, translating vocabularies, coordinating decisions, remains underexplored. This workshop examines LLMs as collaborators, in data-intensive workflows. We take Interdependence Theory as a starting lens to reason about dependence, mutual responsiveness, and shared outcomes in human-LLM interaction, while explicitly reasoning on its fit and considering alternative or complementary frameworks. Through interactive discussions and case-driven activities, we will surface core principles, design considerations, and evaluation criteria (e.g., trust, coordination, equity, transparency) for human-LLM collaboration. Expected outcomes include an initial conceptual framework, high-level guidance for practice, and a forward-looking research agenda to inform the design and assessment of collaborative, responsible LLM-enabled data workflows.