Algorithms in the Stacks: Investigating automated, for-profit diversity audits in public libraries

TL;DR

The paper examines how automated, for-profit collection diversity audits used by U.S. public libraries shape holdings, workflow, and governance. Through a mixed-methods approach (survey of 99 librarians, 14 interviews, GovSpend purchasing data, and vendor materials), it finds that these tools offer speed and benchmarking but tend to flatten diversity into broad categories, rely on proprietary metadata, and deepen vendor dependence. The study highlights concerns about transparency, metadata quality, and misalignment with local community needs, while noting the practical utility of audits for justification and decision support under resource constraints. It concludes with concrete recommendations for transparency, flexibility, open-source alternatives, human-in-the-loop processes, and collective action to ensure that automation serves the public good in an increasingly contested political landscape.

Abstract

Algorithmic systems are increasingly being adopted by cultural heritage institutions like libraries. In this study, we investigate U.S. public libraries' adoption of one specific automated tool -- automated collection diversity audits -- which we see as an illuminating case study for broader trends. Typically developed and sold by commercial book distributors, automated diversity audits aim to evaluate how well library collections reflect demographic and thematic diversity. We investigate how these audits function, whether library workers find them useful, and what is at stake when sensitive, normative decisions about representation are outsourced to automated commercial systems. Our analysis draws on an anonymous survey of U.S. public librarians (n=99), interviews with 14 librarians, a sample of purchasing records, and vendor documentation. We find that many library workers view these tools as convenient, time-saving solutions for assessing and diversifying collections under real and increasing constraints. Yet at the same time, the audits often flatten complex identities into standardized categories, fail to reflect local community needs, and further entrench libraries' infrastructural dependence on vendors. We conclude with recommendations for improving collection diversity audits and reflect on the broader implications for public libraries operating at the intersection of AI adoption, escalating anti-DEI backlash, and politically motivated defunding.

Figures (5)

• Figure 1: A library collection diversity audit report from the vendor Ingram, showing representation of books per diversity category.
• Figure 2: A dashboard featuring results from Baker & Taylor's CollectionHQ DEI Analysis tool, showing representation of books per DEI category.
• Figure 3: Reasons libraries chose to conduct a vendor-provided collection diversity audit, according to survey respondents.
• Figure 4: This figure shows diversity-related purchases for 24 U.S. public libraries, drawn from GovSpend (for any item exceeding $20 and including the words "diversity", "dei", "icurate", or "collectionhq"). Average item price is listed in parenthesis. Diversity audit services were provided by Ingram and Baker & Taylor, and the price of such services cost several thousand dollars. A wider range of organizations provide services that aim to educate library employees on the concept of diversity. Libraries appear to spend more on staff DEI training and consulting than collection audits, averaging around $10,000. However, as Table \ref{['tab:LibA']} and \ref{['tab:LibB']} show, libraries spend more with vendors that provide collection audits overall.
• Figure 5: This figure shows diversity-related purchases from 57 U.S. public libraries. Purchase records are drawn from GovSpend and aggregated by type of seller and purchase; we include average item price in parenthesis. These results reflect any purchases on GovSpend that include "diversity", "dei", "icurate", or "collectionhq" in the item description; include "library" in their agency name; and exceed $20. These figures represent the minimum amount libraries spent during this period.