Part II: AI tools to help with scholarly literature

This is the second of a three-part series on Artificial Intelligence (AI) in biomedicine. Part I introduced readers to the basics of AI and Part III will focus on authorship issues related to AI and publishing.

In the realm of academic research and beyond, AI-powered tools and platforms have become indispensable for scientists and researchers. These tools leverage the power of large language models to assist in various aspects of research, including data analysis, text generation, and predictive modeling.

As AI continues to evolve and improve, so will AI research tools’ capabilities. “The prospective benefits are substantial – from assisting patients in pursuit of improving their health, to providing guidance for clinicians in diagnosis and treatment, to helping researchers find information and develop new ideas to advance their research.”1 However, it is important to note that these tools are meant to augment human intelligence rather than replace it entirely. The expertise and creativity of researchers remain essential in formulating hypotheses, designing experiments, and interpreting results.

While there have been announcements from WashU about AI and its use, see list below [WashU Specific Announcements], we want to provide information how various AI tools can be used to assist researchers in areas of literature review, literature analysis, summaries and presentations. These tools can provide recommendations for relevant literature based on specific research topics or aid in reviewing and summarizing multiple articles.

The following tables provide information on select AI tools that Becker librarians have reviewed and identified that might be of interest to WashU researchers. (Tools listed alphabetically)

Literature Search & Review

Tool Name, URLKey featuresAccess
ASReview lab,• Open source
• Uses active learning to aid in efficient review and screening of articles
• Open source & free software download (runs on Python)
Consensus,• Takes questions in natural language
• Uses Semantic Scholar as a primary database
• Integrates with reference managers
• Unlimited searches, research quality indicators
• 20 AI credits/mth for GPT-4 Summaries, Consensus Meters, & Study Snapshots
Elicit,• Works best for empirical domains
• Mines key information from PDFs and organizes in table format
• Uses Semantic Scholar as a primary database
• 5,000 one-time credits
ResearchRabbit,• “Spotify for Papers”
• Create collections, alerts, and share
• Visualizations showing how articles cited
• Core discovery features are free, accepts donations similar to Wikipedia
Semantic Scholar,• Searches through over 214M papers with filters for journals, authors, pub type and date
• Email alerts
• Integrates with Zotero
• Tools like TLDR (too long; didn’t read) and Highly Influential Citations to help review findings
• Open access, free for all

Literature Mapping / Citation Analysis [start with seed papers, build from those]

Tool Name, URLKey featuresAccess
Connected Papers,• Visualizations of similar papers
• Can add papers for a multi-origin graph
• Database connected to Semantic Scholar Paper Corpus
• Core discovery features, 5 free graphs/mth
Inciteful,• Literature Connector tool -Find connections between 2 papers
• Paper Discovery tool -Discover relevant literature based on paper search
• Data sources are OpenAlex, Semantic Scholar, CrossRef, Open Citations
• Open access, free for all
Litmaps,• Uses Seed (starting paper) to find relevant literature
• Can discover papers related to a topic
• Core features, up to 20 inputs; visualization – 100 articles, 1 Litmap
Scite,• Can tell you where an article has been cited and if citing article affirms or disputes citation
• Integrates with reference manager
• Has dashboards to track trends and groups of papers
• Not free, but offers a short free trial

Article Summaries

Tool Name, URLKey featuresAccess
Explainpaper,• Upload article, highlight sections of article text, AI tool provides summary explanation• Core features free (unlimited highlight explanations & follow-up questions)
Humata,• Built to work with files you upload and generates answers from that data input
• Unlimited files can be summarized quickly
• Highlights citations
• Basic features for 60 pages/mth, also have student rate for paid version
PDFgear Copilot,• Uses natural language to complete tasks with a PDF
• Upload files and get summary of information
• Free software download
Scholarcy,• Good for research articles
• Copy and paste or upload pdf
• Organized sections for easier review
• Comparative analysis
• Embedded links to access articles or additional info
• Ability to store notes with summarized article
• Core features free (3 summaries/day; export flashcards individually)
SciSpace,  • AI platform to help with discover, analyzing and writing scientific literature
• Drag and drop file feature
• Copilot answers from standard model, limited messages, review searches

Niche/Specialty Tools

Tool Name, URLKey featuresAccess
Journal Author/Name Estimator (JANE),• Suggests journals for manuscripts based on abstract or keywords
• Compare your document to millions of documents in PubMed to find the best matching journals, authors or articles
• Open access, free for all
Lumen5,• Video creation platform using machine learning, automate tasks to help users create quality videos with minimal time and effort
• Drag and drop interface
• Create videos from blogs
• Templates for video images and styles
• Limited to 5 videos/mth, 720p video resolution, Lumen5 watermark
NightCafe,• AI image generator, can do text-to-image or image-to-image
• User friendly interface
• Multiple modes and styles
• Limited to 5 daily credits
WashU GPT,• Secure sandbox where you can use sensitive data and WashU intellectual property• Free for all WashU employees, sign-in with WUSTL key


WashU-Specific Announcements

  1. Hersh W. Translational AI: A Necessity and Opportunity for Biomedical Informatics and Data Science. Musings from the Mezzanine. February 7, 2024. Date accessed March 25, 2024.