Artificial Intelligence in Systematic Reviews: Uncharted Waters for Librarians Public Deposited

Downloadable Content

Download PDF
Last Modified
  • May 9, 2019
Creator
  • Moreton, Elizabeth
    • ORCID: orcid.org/0000-0003-1769-7134
    • Affiliation: University of North Carolina at Chapel Hill. Health Sciences Library
Abstract
  • Background: For over 10 years, informatics journals have published articles about the potential for artificial intelligence (AI) tools to automate parts of the time-consuming and labor-intensive systematic review process. Despite this vision, automation of the systematic review process is still uncommon today for many research teams. This study will investigate why these potentially time-saving tools have not been incorporated into librarians’ systematic review workflow, especially given the recent surge in systematic review requests. Methods: This study will involve a review of the literature which will explore the proposed uses of AI to facilitate the systematic review process. In particular, the review will identify tools that can incorporate AI into the systematic review screening process, as well as resources and best practices for the use of these tools. Finally, the study will examine the facilitators and barriers to librarians’ adoption of AI tools in systematic reviews. Results: The researchers will present key findings from the literature review. They will explain types of AI and available options for librarians’ use to expedite the systematic review screening process. The researchers will share a list of available tools and recommendations for librarians to begin to incorporate AI into their systematic review workflow.Conclusions: Explaining the history, uses, and potential benefits and challenges of AI in systematic reviews may help librarians understand how best to incorporate these tools into their expert searching services.
Date of publication
Keyword
Resource type
Conference name
  • Mid Atlantic Chapter of the Medical Library Association Annual Meeting
Language
Parents:

This work has no parents.

Items