AI in Library Databases
AI assistants are increasingly embedded in tools and technologies that we use every day (e.g. Google's AI Overviews), and library information resources are no exception: Over the past two years we have seen a rapid rise in AI assistants built into content management systems and familiar academic databases like Web of Science and ProQuest. While AI assistants in discovery tools are still considered an add-on to existing products, it is likely that these AI elements will eventually become an inseparable part of the product themselves.
As of May 2025, the Mansfield Library has activated most available AI assistant integrations linked to current library databases and will continue to do so as vendors develop additional tools. This page contains a brief overview of AI assistant features and capabilities in UM's library databases. Faculty librarians at the Mansfield Library are actively engaged in research seeking to better understand how use of these AI assistants impacts students' information literacy and research skills. Please feel free to reach out if you are an instructor with questions or concerns!
You can see which UM databases include generate AI tools by navigating to the A-Z Database list and selecting "Includes generative AI tools" in the "Content Types" dropdown menu:
How Can I Identify an AI Tool?
Look for the sparkle emoji:
This symbol has gained traction as the de facto symbol for AI tools, particularly when embedded into another tool or system. Many (though not all) of UM's academic databases use some version of the sparkle emoji to indicate access to an embedded AI tool. Here is an example:
Types of AI Integrations
AI tools integrated into academic databases provide different functionalities for various steps in the research process. The following descriptions are general descriptions of AI functionalities and are not specific to vendors or databases. Note that each feature may operate slightly differently or be found in different places from database to database.
Summaries and Key Points
These integrations offer concise summaries of articles and other text-based content, often extracting 2-5 key ideas from each document.
Natural Language Search Mode
Natural language search mode allows users to query using everyday language (instead of using keywords and Boolean operators), enabling you to search in a more conversational manner.
Research Assistant
Research assistant tools often combine functionalities, offering research support through natural language searches, generated text responses to search queries, citations for relevant resources, and suggestions for similar research topics.
