Guides: Generative AI and Research: AI Research Platform Tools

Introduction to AI Research Platforms

AI-powered research tools have become increasingly important in academic research, offering other ways of doing research beside the traditional library databases. Tools like Perplexity, Scite, Elicit, Keenious, Consensus, ResearchRabbit, and OpenAlex provide various features, including citation analysis, literature discovery, and evidence-based insights. While these tools offer significant advantages in streamlining tasks such as literature reviews, identifying relevant papers, and tracking citation networks, they also have limitations, such as coverage gaps, reliance on metadata, and the need for validation of the sources.

AI Research Platforms

Perplexity
"Real time search engine, cites sources, emphasis on accuracy.
Search focus: narrow down the sources used to generate the response (can search just within academic papers, or within Youtube, for example).
Copilot features asks follow up questions to narrow down your question to get more accurate results."
Elicit
"A research assistant using LLMs to automate lit review, primarily through synthesis and extraction. Uses semantic similarity rather than keywords for search. Generates custom summaries based on your query."
Keenious
"Analyzes inputs (text, URL, pdf) to find relevant scholarly articles and related topics. Data for results coming from OpenAlex. Word and Google Doc extensions available. Use filters to sort further through results.
Highlight search–highlight key parts of inputted document and search will be tailored accordingly. One key use case for both researchers and students is using the tool to check your literature review has not missed any important literature on the given topic."
Scite
"“Helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. Scite Assistant uses generative AI to answer questions by querying their citation index. Chrome extension.
Enterprise version can track how research from across the institution is being cited. Zotero plugin available. API available."
Consensus
"AI-powered scientific search engine” to summarize areas of consensus in academic research. Step by step instructions on how Consensus works. Save searches and individual citations. GPT-4-powered summaries. Summarizes relevant findings when you search. Provides list of specific citations used to generate summary Includes only peer-reviewed sources."
OpenAlex
"At the heart of OpenAlex is our dataset—a catalog of works. A work is any sort of scholarly output. A research article is one kind of work, but there are others such as datasets, books, and dissertations. We keep track of these works—their titles (and abstracts and full text in many cases), when they were created, etc. But that's not all we do. We also keep track of the connections between these works, finding associations through things like journals, authors, institutional affiliations, citations, concepts, and funders." [website].
ResearchRabbit
“Search for papers and authors, monitor new literature, visualize research landscapes, and collaborate with colleagues." Mission statement.
Search powered by PubMed and Semantic Scholar.
Create “collections” of papers and receive recommendations (“Spotify for papers”), visualize networks of papers and authorship, collaborate on collections. Emphasis on visualization-based searching. Zotero integration."
Undermind
"AI-powered scientific research assistant. Provides list of relevant articles for conversational search query + summary.
Draws from Semantic Scholar database; currently searching abstracts and metadata but plan to incorporate full text discovery in the future.
Chat with research assistant until it fully understands what you are looking for to perform most accurate search.
Gives percentage of how relevant article is to your search. Also provides short summary next to each result of why it is relevant.
Estimates percentage of existing papers on the topic it has searched, with option to expand search."
Semantic Scholar
"Free, AI-driven search and discovery tools, and open resources for the global research community. Provides 1 sentence TLDR for journal articles. Users can create unique libraries and receive AI informed recommendations and notifications of new papers.
API access available. Institutional log-ins available."

Quotations are taken from the Ithaka S+R Generative AI Product Tracker.

Deciphering the AI Research Platform Maze: A Comparative Analysis

Charles Gallagher and Kevin Gunn, librarians at The Catholic University of America Libraries, gave this workshop to their colleagues at The Washington Research Library Consortium Annual Meeting on May 22, 2024.

Testing AI Tools

As with any tool used in digital scholarship, you should spend some time researching the best tool to answer your research questions. Two excellent resources include:

Temple University Libraries has created an AI Tools Report Card assessment rubric for asking questions in evaluating AI-powered research tools.

ITHAKA S+R created the Generative AI Product Tracker that lists generative AI products marketed to faculty and students for teaching, learning, or research. This list is updated regularly. Currently 63 pages long!

Rubric for Evaluating AI Tools

Librarians at the State University of New York at Oswego have created a rubric for evaluating AI research tools.

Caico, Marissa; Harris, Laura; O'Shea, Sarah; Mitchell, Emily. 2024. Evaluative Information Literacy Rubric for AI Tools.