Search

Diversity audits reveal the extent to which library collections incorporate the topics, perspectives, authors, characters, and narratives associated with underrepresented or marginalized groups. They can help us evaluate whether the collection is aligned with the user community, pinpoint the specific areas where improvement is needed, establish goals and timelines, measure progress toward those goals, demonstrate that deficiencies in the collection are being addressed, and highlight the need for additional resources. This guide reviews the scholarly and professional literature on diversity audits of academic library book collections. It discusses how concepts such as diverse groups and diverse books can be operationalized; describes three methods of conducting the diversity audit – the catalog search method, the checklist method, and the book inspection method; considers various benchmarks or targets that may be adopted; and explains how the results of the analysis can be used to build more diverse library collections. The review concludes by presenting three fundamental questions that explore the limits of the audit methods currently in use, highlight underlying issues that librarians may want to consider, and suggest avenues for further research and discussion.

This study evaluates the accuracy of 16 publicly available AI text detectors in discriminating between AI-generated and human-generated writing. The evaluated documents include 42 undergraduate essays generated by ChatGPT-3.5, 42 generated by ChatGPT-4, and 42 written by students in a first-year composition course without the use of AI. Each detector’s performance was assessed with regard to its overall accuracy, its accuracy with each type of document, its decisiveness (the relative number of uncertain responses), the number of false positives (human-generated papers designated as AI by the detector), and the number of false negatives (AI-generated papers designated as human). Three detectors – Copyleaks, TurnItIn, and Originality.ai – have high accuracy with all three sets of documents. Although most of the other 13 detectors can distinguish between GPT-3.5 papers and human-generated papers with reasonably high accuracy, they are generally ineffective at distinguishing between GPT-4 papers and those written by undergraduate students. Overall, the detectors that require registration and payment are only slightly more accurate than the others.