Skip to Main Content Saint Mary's Libraries

Note on AI Research Tools

While these tools offer valuable features for streamlining literature reviews and assessing research quality, it is crucial for researchers to maintain critical thinking, fact-checking, and oversight. Overreliance on AI can lead to issues like propagation of bias or errors and missed nuance or context. The key is to leverage the strengths of the tools while complementing them with domain expertise and ethical research practices.

Connected Papers

Connected Papers is a visual tool that helps researchers explore and discover relevant academic papers based on citation analysis and similarity measures. It generates interactive graphs where papers are represented as nodes, with closely related papers clustered together and connected by links.

Best used for:

  • Conducting literature reviews by identifying core papers and their connections
  • Discovering recent research and seminal works in a specific field
  • Visualizing citation networks, author collaborations, and research trends
  • Identifying potential research gaps or under-explored areas

Limitations include:

  • Limited coverage in certain disciplines (especially humanities, some social sciences, and book-oriented disciplines)
  • Accuracy depends on the underlying algorithms for measuring paper similarity
  • User interface can be challenging to navigate, especially for large paper collections

Consensus

Consensus uses AI to find answers in research papers. It uses the Semantic Scholar database, which includes over 200 million papers across all domains of science. Searches produce both summaries of the literature and a list of highly cited articles for further reading. Since it is pulling directly from academic databases, it does not have the source hallucination problem of other AI tools. 

Best used for:

  • Generating literature review matrices, summaries, and insights using AI
  • Tracking new research papers based on topic and getting personalized recommendations

Limitations include:

  • Limited coverage in certain disciplines (especially humanities, some social sciences, and book-oriented disciplines)
  • May struggle with niche or complex topics
  • Requires careful verification for accuracy

Elicit

Elicit is an AI tool to find 'seed articles' to mine for keywords/subject headings.  When you enter a question, it returns alternate questions that can lead to further "seed" articles. Elicit uses language models to help you automate research workflows, like parts of literature review. Elicit can find relevant papers without perfect keyword match, summarize takeaways from the paper specific to your question, and extract key information from the papers. While answering questions with research is the main focus of Elicit, there are also other research tasks that help with brainstorming, summarization, and text classification.

Best used for:

  • Conducting literature reviews by filtering through documents to find the most relevant studies
  • Generating novel research hypotheses by analyzing existing literature to identify under-researched areas
  • Quickly summarizing core findings and methodologies of research papers to assess relevance

Limitations include:

  • Reliance on AI-generated summarises and insights, which may miss nuance or misinterpret context
  • Potential bias or inaccuracy propogated from the AI training data
  • Inability to replace critical evaluation of sources and original thinking by human researchers

ResearchRabbit

Research Rabbit is a discovery engine. It bills itself as "the Spotify of papers". It is tool for exploring new areas and writing literature reviews. It allows you to uncover new and classic papers related to your area of research. It also shows links between authors - making maps of collaborations and related works. Research Rabbit will also send you a weekly digest of new papers that it has discovered that may be relevant to your research. 

Best used for:

  • Conducting literature reviews by identifying relevant papers and exploring connections
  • Discovering new research and authors in a specific field or topic
  • Visualizing citation networks and author collaborations
  • Identifying potential research gaps or unexplored areas

Limitations include:

  • May not cover all publications, leading to potential gaps
  • Requires time to understand and effectively use all features
  • Relies on AI for recommendations, which might not always align with human judgment or miss nuanced connections

Semantic Scholar

Semantic Scholar is a free, AI-powered research tool for scientific literature, based at the Allen Institute for AI. More than 200 million papers are sourced from partners such as PubMed, Springer Nature, Taylor & Francis, SAGE, Wiley, ACM, IEEE, arXiv, and Unpaywall. Access is free at semanticscholar.org. No login is required. To save papers to your library and receive custom alerts, sign in with your institutional login, a username and password, or with your Google, Twitter, or Facebook accounts.

Best used for:

  • Discovering relevant research papers using AI and semantic analysis
  • Getting summarized key points and highlighted insights from papers
  • Exploring similar papers and influential citations

Limitations include:

  • Limited coverage in certain disciplines (especially humanities, some social sciences, and book-oriented disciplines)
  • Occassional inaccuracies in analysis and associations

Scite.ai

Scite.ai is an AI-powered research platform that helps users analyze and discover relevant research publications. It uses natural language processing techniques and deep learning algorithms to analyze the text of scientific papers and their citation contexts. 

Best used for:

  • Helping the researcher assess the credibility and reliability of research papers through citation analysis
  • Quickly identifying influential studies based on how they have been cited
  • Providing insights and highlighting important citation information

Limitations include:

  • Full access costs money
  • The primary focus on citation analysis may not capture the full context or nuance of a study
  • Inability to generate original context, as it is an "extractive" tool

Note on Content

This page was written with the help of the AI tool Perplexity and fact checked and edited by librarians.