Integration of AI assistance into exploratory data analysis

Determine how to integrate AI assistance into exploratory analysis of unstructured data in a manner that improves analytical efficiency while preserving user control.

Background

The paper presents exploratory analysis of unstructured data as a process in which users gradually construct knowledge structures while examining data. Because the possible organization strategies are substantially different, AI assistance must be integrated in ways that support users without imposing an inappropriate taxonomy, category system, or classification scheme. The authors explicitly identify the ultimate integration of AI assistance, while retaining user control, as unresolved.

References

It is also an open question how to ultimately integrate AI assistance for more efficient data analysis while leaving the user in control.

— Exploratory Unstructured Data Analysis: A Formative Study and Implications for Human-AI Collaboration  (2609.03678 - Eschner et al., 3 Sep 2026) in Section 2.1, “Knowledge Organization”

The questions this leaves open are the ones the design cannot settle on its own: whether the intent graph's progression aligns with the direction an analyst is actually pursuing, whether the knowledge graph meaningfully augments their understanding of the data domain and the intent graph branching, and whether the two graphs are used as scaffolds for the analyst's own reasoning.

— Intent Graph: Navigating the Analytical Reasoning Space for Exploratory Data Analysis  (2610.11025 - Yang et al., 8 Oct 2026) in Section 4, “Limitations and Future Work” (\ref{sec:dag-future})