Skip to Main Content

Artificial Intelligence (AI) Guide

Gather & Evaluate Overview

Topics
  • Identifying and Evaluating High-Quality Evidence
  • Screening Articles for Inclusion/Exclusion
  • Citation Management
  • Data Extraction and Analysis
Limitations & Questions
  • Be careful about uploaded full-text articles into an AI tool. You may violate copyright or license agreements. Also, be careful about uploading your original research data/manuscript. Unless it is an closed model or there is an enterprise agreement, there is always the possibility the company can retain your data and use it to train their tool.
  • Many AI tools are limited to open-access (or publisher-agreed) content. They cannot often access paywalled content, meaning you will miss some full-text articles if you rely on external AI tools exclusively.
  • Do not become overdependent on AI. Apply your creativity and critical thinking. AI models may lead to superficial synthesis and analysis due to a lack of contextual understanding and nuance.

Gather & Evaluate Tools

Why These, Not Those?
  • Many AI tools are limited to open-access (or publisher agreed) content. They often cannot access paywalled content, meaning you will miss some full-text articles. We provide tools that sync to the library's journal holdings and can access a much wider breadth of content.
  • We recommend tools that have been tested and accepted in the field of evidence synthesis, such as Covidence.
  • You can receive more support for the use of these tools through consultation with a librarian or vendor support lines.

Note: All AI can hallucinate, provide poor quality outputs, or oversimplify complex topics. Apply critical analysis to any output.

Use the ROBOT test to evaluate AI tools.

 

Please review the Discover page to find out which bibliographic databases incorporate article summarizers.

Screening & Data Extraction

Use these tools to screen articles for inclusion/exclusion, manage critical appraisal, and extract data from articles.

Covidence assists with streamlining collaborative title and abstract screening, full-text review, risk-of-bias assessment, and data extraction.

Elements of machine learning are integrated into Covidence:

  • RCT Classifier – Identifies and auto-tags RCTs and non-RCTS. There is an option to automatically exclude non-RCTs from screening consideration when setting up a review (use carefully).
  • Relevancy Sorting – Sorts your study list by the most relevant articles based on previous screening decisions.
  • Auto Data Extraction – In some cases, it will automatically extract limited metadata as suggestions. These suggestions appear in yellow in the extraction template. You can accept or reject these suggestions.
Citation Management

Use these tools to manage your research citations and generate bibliographies in your manuscripts.

EndNote 2025 is now available for download on Windows AND Mac devices.

Note: The download link requires that you are either on campus (i.e., using the UTSW wireless network from your laptop) or using VPN remotely. This software is not to be installed on CMC or PHHS devices.

  • Key Takeaway – An AI-generated summary of the key findings of the article and topics discussed. It will automatically generate when a PDF is uploaded. Select PDF in the reference panel to access.
  • Full-Text Pathways – You can set up pathways to import PDFs for citations into EndNote.
  • Research Assistant – Chat and ask questions about PDFs in your library, quickly summarizing or translating content.
Finding Full Text

Use these tools to find full-text PDFs for articles. They do not guarantee access to the full text, as they are connected to library holdings, but they can provide quick access points.