Survey of Use of Bard, Bing & ChatGPT for Academic Library Cataloging

Survey of Use of Bard, Bing & ChatGPT for Academic Library Cataloging

Looking at the entire sample, over half (53.85%) have not used the AI tools for cataloging. Only a minority finds it unreliable (3.85%) while 23.08% find it occasionally unreliable but useful, and 11.54% find it sufficiently reliable that they can

edit any issues out and find it very useful. None of the participants find it extraordinarily useful.

When breaking down by tuition costs, more expensive institutions ($40,000+) show a relatively high percentage (60%) of respondents who have not used these tools, while

the cheaper ones ($6,000 or less) show a majority (71.43%) in the same category. Institutions with tuitions from $12,000 to $40,000 seem to show more varied

This study presents data from 23 colleges and universities about how they are using various artificial intelligence applications in academic library cataloging.  The report gives separate data sets for Bard, AI-enabled Bing and ChatGPT, defining the extent and kind of use.  The study also gives data for use of these programs for particular applications, such as for  subject headings, descriptions, indexing, cataloging training and other uses.  The data also explores the impact of use of these applications on other cataloging tools such as the Library of Congress Catalogers Desktop Developers Tool, MarcEdit and OCLC’s Connexion.  In addition, the report details exactly how much time catalogers are spending on AI tools, their level of overall satisfaction with them, and their plans for the future.