Skip to Main Content

A Guide to AI for Gonzaga Faculty

Google NotebookLM

  • NotebookLM is an AI-powered research and note-taking assistant developed by Google, built on their Gemini LLM. Its core distinction is that it operates primarily with the content you provide, in the form of uploaded documents and files.
  • NotebookLM allows a researcher to upload a set of documents or data and interrogate it. NotebookLM will provide links to document and portion it is referring to in its answers. NotebookLM can still hallucinate, especially when dealing with very long files, but since it references its answers it is easier to check its outputs 
  • Can generate summaries, audio overviews, podcasts, briefings, study guides, quizzes, timelines, and more
  • The free version allows 50 uploaded documents of up to 500,000 words each, and 50 chat interactions and three audio generations daily. The paid tier raises those limits to 300 documents, 500 daily chat interactions and 20 daily audio generations

Consensus.app

  • A sciences-focused research database and search tool
  • Unlike traditional keyword-based search engines, Consensus uses LLM-based natural language processing to “understand” the meaning and context of queries, providing relevant results without the need for complex search techniques
  • Searches must be phrased as literal queries, i.e. questions, such as “Does the death penalty reduce crime rates?" or "Does a lack of sleep increase the risk of Alzheimer's?"
  • Presents a short summary of the answer and the top twenty articles examining the question, taking into account both the relevance to the question and the impact of the journals. Every answer is thus grounded in and links back to legitimate research
  • The corpus which is searched includes over 200 million peer-reviewed articles drawn from Semantic Scholar, OpenAlex, and Consensus’ own additional dataset
  • Has a “Consensus Meter” which uses AI to indicate the level of agreement or disagreement to a yes/no question
  • Can summarize and extract key features of individual papers, and allows the user to ask questions about individual papers
  • Hallucination is still possible, but only as a misinterpretation/misunderstanding of existing and linked data, not making something up from whole-cloth
  • Has a free tier with search and feature limitations, and a paid tier which raises these limits

Perplexity

  • An LLM-powered search engine for the general internet
  • Provides links back to sources
  • Focuses on finding factual answers, rather than conversation or creative works
  • Can still hallucinate
Accessibility | Proxy Logout