Note: OpenAI leads the general AI space, but AI companies are developing deep research tools and experimenting with AI-powered academic searching. Perhaps you have faculty or students asking you to present these tools. For more information, see Which companies are behind AI search tools?.
Note: This open textbook (or wiki channel) is intended to help librarians and other information professionals learn about AI. It is not, in itself, meant to be seen as promotion of AI. If anything, the goal is harms mitigation or harms reduction.
Definition
"AI-powered searching" refers to systems with AI features and/or supported by machine learning and/or natural language processing to enhance the location, collection and synthesis of papers. This type of searching is a type of AI-powered discovery or AI search and synthesis.
As this search area evolves, watch for AI tools to supplement traditional lexical "word-based" approaches (and "indexed article" search systems) by shifting to understanding the context and meaning of queries, extracting relevant data from papers, and providing structured insights to support research tasks such as literature reviews, systematic reviews, and knowledge synthesis (KS). This type of searching overlaps with agentic searching and deep research technologies. SeeArtificial intelligence (AI) glossary of terms.
As more and more of our traditional bibliographic databases incorporate natural language searching with AI chatbots and summaries, librarians need a way to respond and teach these features. With changes and development happening every day, each librarian seems to be approaching this situation differently. Any discussion of this topic should include the strategic, systematic and ethical issues associated with using AI tools in knowledge synthesis (KS). This entry is intended to help librarians and other information professionals learn about AI. It is not, in itself, meant to be seen as promotion of AI.
What are some key characteristics of AI-powered search tools?
Here are some key characteristics or underlying technologies to watch for that support AI-powered searching:
Access to broad corpus using retrieval augmented generation (RAG): many tools, including Semantic Scholar and Elicit.com, use generative searching of a large number of papers (e.g., Semantic Scholar’s 200+ million papers) to ensure comprehensive literature coverage.
Semantic searching: while traditional databases rely on exact keyword matches, AI tools such as Elicit.com, Semantic Scholar, and Undermind.ai use semantic understanding to interpret natural language queries and find conceptually relevant / related papers.
Data extraction and summarization:Elicit.com and Scite.ai automatically extract key information (e.g., study methods, findings, sample sizes) from papers and present it in structured formats, such as tables or summaries, reducing manual effort.
Citation analysis:Scite.ai analyzes how papers are cited (e.g., supporting or contradicting), provides insights into reliability and impact.
Automation of research workflow:Otto-SR and Elicit.com streamline systematic reviews by automating screening, data extraction, and synthesis, potentially reducing research time by up to 80%.
Specialized niche tools:PubMed.ai focuses on biomedical literature, while Undermind.ai emphasizes in-depth, iterative searches for complex research questions, tailoring results to specific fields.
Vector search is used to retrieve documents, articles, web pages or other textual content based on their similarities to queries. Vector searching enables users to find relevant information even when the exact terms used in the query are not present in the documents. The first step in vector searching is translating text to vectors (or numbers) and processing them through a large language model. VS is a type of AI system trained on vast amounts of text (for example, research papers, books, and web content). During training, the model learns how words appear together and in what contexts. Over time, it builds a kind of map of language, where meanings cluster naturally. In medicine, words that often appear in similar contexts, such as doctor and physician, end up close together in this semantic map. Words that rarely co-occur or belong to very different contexts, like insulin and wheelchair, are far apart.
Examples of AI-Powered Academic Search Tools
Elicit.com automates literature reviews by searching 125 million papers, summarizing findings, and extracting data into tables.
Scite.ai analyzes citation contexts to assess the reliability of claims and supports researchers by highlighting supporting or contradictory citations.
Semantic Scholar offers free AI search across 200+ million papers, extracting insights and identifying connections to aid in scientific discovery.
SciSpace AI-powered platform or "co-pilot" designed to assist researchers, students, and academics in streamlining their research and writing.
Otto-SR specializes in automating systematic reviews, focusing on screening and data extraction for large-scale research projects.
PubMed.ai tailored for biomedical research and uses AI to enhance search precision within PubMed.
Undermind.ai provides in-depth searches, generating overviews 10-50 times more effectively than traditional methods like Google Scholar.
Presentation
Note:Julie Glanville considers whether general purpose AI tools and research-focused AI tools can aid in search strategy development for systematic reviews; also, how AI tools of various types might assist with search planning and study identification. Includes discussion of tools.
Search, Screen and Extract using AI
Systematic reviews are essential for synthesizing research evidence, but the process is labor-intensive and time-consuming, particularly during the searching and study selection phases. Traditionally, these two phases of the SR are performed manually by human reviewers (librarians and researchers) to ensure thoroughness and quality. While tools such as Semantic Scholar and Scite.ai appear to be focussed on searching, Elicit.com and Undermind.ai are designed to accomplish more than searching, including screening and extracting data. The goal for new AI search tools is to streamline the process from searching through review generation.
López-Pineda et al, 2025 concluded that "...integrating AI into SRs could transform the management of large datasets by significantly reducing the time and effort required. These advancements suggest that AI can be a valuable complementary tool, offering improved automation without compromising quality, thereby optimizing the study selection process and potentially elevating the overall effectiveness of systematic reviews."
PubTator 3.0 (https://www.ncbi.nlm.nih.gov/research/pubtator3/) is a biomedical literature resource using state-of-the-art AI techniques to offer semantic and relation searches for key concepts like proteins, genetic variants, diseases and chemicals. It currently provides over one billion entity and relation annotations across approximately 36 million PubMed abstracts and 6 million full-text articles from the PMC open access subset, updated weekly.
LitSense 2.0 (https://www.ncbi.nlm.nih.gov/research/litsense2/) is an advanced biomedical search system enhanced with dense vector semantic retrieval, designed for accessing literature on sentence and paragraph levels. It provides unified access to 38 million PubMed abstracts and 6.6 million full-length articles in the PubMed Central (PMC) Open Access subset, encompassing 1.4 billion sentences and ∼300 million paragraphs, and is updated weekly.
Disclaimer
Note: Please use your critical reading skills while reading entries. No warranties, implied or actual, are granted for any health or medical search or AI information obtained while using these pages. Check with your librarian for more contextual, accurate information.
