Skip to Main Content
Reed Library

AI Research Tools

Elicit

Overview

Elicit is an AI research assistant created by Ought, though it is now an independent business. Unlike Semantic Scholar and Research Rabbit, which use traditional machine learning algorithms, Elicit is powered by large language models (LLMs) similar to ChatGPT. This allows Elicit to "understand" natural language questions, read and comprehend research papers, and extract specific information from studies. The LLM technology enables it to not just find papers, but to actually "read" them, pull out relevant data points, methodologies, and findings, and "chat" with documents.

REMEMBER: As Elicit is powered by LLMs, it's possible that outputs might be "hallucinated" in whole or in part. This means that they may not be accurate. See the AI Literacy Guide for more.

Access: https://elicit.org/ (need to create a free profile)

Key Features

Direct Answer Synthesis

  • Ask research questions and get synthesized answers from multiple papers
  • LLM reads and combines findings across studies
  • Provides citations for each claim

Paper Summarization

  • Generates custom summaries of individual papers
  • Focuses on aspects relevant to your research question
  • Goes beyond abstracts to analyze full paper content

Automated Data Extraction

  • LLM "reads" papers to extract specific information
  • Pulls out sample sizes, methods, outcomes, and findings
  • Creates structured data from unstructured text
  • Export extracted data as CSV files

Research Question Breakdown

  • Analyzes your question to identify key concepts
  • Suggests related questions you might explore
  • Shows which aspects are well-studied vs. gaps in literature
  • Helps refine broad questions into researchable components

Video Tutorial

Best For

  • Those who want LLM functions integrated into the resource discovery process
  • Advanced researchers with a thorough understanding of the pros and cons of LLMs in research
  • Those who have time to check results against traditional methods