Reflections on tests of AI-search tools in the academic search process

Abstract

Academic search literacy and artificial intelligence (AI) search are a strategic focus of future researcher support strategies at the Royal Library and affiliated university libraries in Denmark. The ambition is to integrate AI search tools in teaching and services that support literature seeking and hence improve the efficiency of the academic search process.
 This paper is a reflection on the results of tests of AI powered search tools conducted at the Royal Library, Denmark. Information specialists and researchers took part in think-aloud tests, Hackathon and expert quality assessment in the period of April 2021 – February 2022.
 Our observations point to a) opportunities in AI search to let go of traditional approaches to search and broaden our perception of how an academic search can be conducted, b) potential of AI search interfaces to challenge ones preconception and reduce cognitive bias in the search process, c) concerns about the quality of scientific literature AI search identifies, and d) barriers that need to be overcome before practical implementation of Ai search tools become viable in academic searches.
 More research is needed to show the usefulness of AI search tools in systematic search and guidance on how to report the use of them