AI & GPT: AI for Research

What AI is useful for (and what's it bad at) as a Writing Assistant

AI is great for....

• Finding related papers in databases
• Proof-reading and line-editing
• Summarizing, creating abstracts, or otherwise distilling content into shorter, easier-to-read language. [note: this should be treated like any translation or interpretation, in which nuance or emphasis may be lost or misplaced!]
• Writing code and frameworks that have already been done a hundred times before (like the collapsible accordion headers in this guide)

AI is not good for...

• Writing a paper from a prompt
• Copy-editing
• Speculating or future-thinking
• Analysis and insight
• Creating novel code for a brand-new use case

______________________________________________

Articles on how to use AI tools for research

Using Generative AI (Deakin)

"Tech is a massive part of being a student and researcher. It connects to many aspects of being an active learner and explorer of knowledge. Generative AI is a type of artificial intelligence (AI) that uses machine learning algorithms to produce or remix content. This content output can be text, imagery, audio, code or other formats.  Generative Artificial Intelligence (AI) tools can be used in many ways in your study or research - to synthesize, create, refine, test, inspire or modify.  This guide provides information about using generative AI tools in an ethical, creative and evaluative way."

Selected AI-Based Literature Review Tools (TAMU)

Overview of AI tools that can be used to begin a literature review.

Using AI Tools in Your Research (Northwestern)

Describes text-based Generative AI tools like ChatGPT, and using them to generate suggestions for keywords for searching. Also lists several proprietary and free research databases with AI features, including Elicit.

PLEASE DO NOT UPLOAD COPYRIGHTED OR PRIVILEGED INFORMATION INTO ANY AI TOOL.

All of the following are considered "open" and not secure.

Beware "hallucinations" with all generative AI tools.

• Consensus AI
  Consensus is an academic search engine powered by AI. They use LLMs and vector search technology to surface the most relevant papers. They synthesize both topic-level and paper-level insights. Its dataset of over 200 million papers was made possible by a partnership with Semantic Scholar.
• Elicit
  Elicit is an AI-powered academic search engine trained on a knowledge base of over 125 million academic papers via the Semantic Scholar database. Summarizes papers, extracts data from PDFs, and identifies concepts. Requires an account.
• Inciteful
  Create a network of academic papers, or input two papers and analyze their relationships.
• Keenious
  By uploading a PDF or copy/pasting text, you can find relevant articles via the Open Alex dataset. It is known for having good privacy protection. That said, please do not upload library materials into Keenious.
• Litmaps
  Freemium tool that uses open access metadata from Crossref, Semantic Scholar, and Open Alex to provide visualization and monitoring tools.
• Research Rabbit
  Research Rabbit is a visual citation-mapping tool that uses PubMed and Semantic Scholar databases. Requires an account.
• SemanticScholar
  Semantic Scholar is a database that includes over 200 million publications from all fields of science. AI features: craft a one-sentence summary (TL;DR); "ask the paper questions;" and a research feed where your ratings of previous work generates recommendations of continued reading.
• SciSpace
  Chat with a PDF to learn more about its contents.