A 2010s-Era MLS Library Reference Assignment Meets 2025-Era AI: An Assessment of the Suitability of AI Tools to Support Scholarly Research

Abstract

This research article presents library reference interview instruction taught to Masters in Library and Information Science students during the 2010s. Information provided in lectures and practical labs is summarized, before a corresponding class assignment is described and explained. The assignment tasked students with researching and summarizing three topics (“Triples”) to assist PhD researchers in the context of an academic library research interview. This research revisits the assignment by conducting a new experiment to assess the performance of six current artificial intelligence (AI) tools when responding to this task. The experiment results report AI tools’ ability to summarize complex topics with reasonable accuracy, breadth, and depth. With only two partial exceptions, however, the provision of accurate and relevant scholarly references remains a predominant and significant deficiency. We conclude, therefore, that at present students should be discouraged from overly relying on AI tools for research assistance. Areas of potential future improvement for AI tools are both identified and encouraged, and a collaborative human/AI partnership is advocated for both the present and near-future. This research attempts to honor the work of Professor Alan Poulter, Department of Computer and Information Sciences, University of Strathclyde, and is dedicated to his memory.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 1 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.