Research Guides: Scholarly Publishing: Artificial Intelligence and Scholarly Communications

Using AI for Research

"AI (Artificial Intelligence) refers to the ability of machines and computers to perform tasks that would normally require human intelligence. These tasks include things like recognizing patterns and making predictions. Ultimately, that’s not magic; it’s math." (Monahan).

There is increasing interest and concern about AI being used for research and writing purposes. While AI can be a helpful tool, it is useful to know what these tools are built for, who builds them, and what their limits are. This guide seeks to define what the AI research tools are designed to do and provide tips on how to use, and not to use, a sampling of available research tools.

Monahan K. Artificial Intelligence, Explained. Carnegie Melon University Heinz College. Accessed December 13, 2024. https://www.heinz.cmu.edu/media/2023/July/artificial-intelligence-explained

Definitions

AI: Artificial Intelligence (AI) is when a machine performs tasks that usually require human intelligence, such as programming, or reading and writing human language.

Generative AI: A type of AI that creates new, original content, such as images, text, or music, based on patterns and structures learned from existing data" (Cornell CTI). In research, this generally refers to writing tools.

LLM: A large language model (LLM) is a type of AI that generates human language responses based on prompts supplied by a human user. It can answer questions, summarize text, provide help with programming, and translate text from one language into another, among other tasks. Examples include Claude, CoPilot, and ChatGPT (University of Arizona Libraries).

Semantic AI Tools: Use LLMs (such as ChatGPT) to search the contents of papers. Less reliable for research.

Citation Tools: Use AI to find connections between papers via citations, and connected citations. More reliable for research.

Cornell University Center for Teaching Innovation. Generative Artificial Intelligence. Access December 20, 2024. https://teaching.cornell.edu/generative-artificial-intelligence

University of Arizona Libraries. What is a Large Language Model (LLM)? Accessed December 20, 2024. https://ask.library.arizona.edu/faq/407985

AI Platforms

ClaudeAI
Claude is a semantic AI tool is one of the more responsible and ethical LLM's available, good for helping to generate ideas and help with the writing process. Includes a free version and a paid model.
LitMaps
LitMaps is a citation AI tool that creates interactive literature maps. Run by scientific researchers, the model presents collections of articles that match research topics searched for by the user. It does not attempt to summarize the article, as some AI platforms do, but presents links to related research in a visually-appealing way. Includes a free version and a paid model.
Microsoft Co-Pilot
Co-Pilot is a semantic AI chat bot built with ChatGPT that can answer simple and complex questions, provide summaries of content, and generate images. Access provided via GW.

Quick reference table

AI tool	what the tool is designed to be	what the tool does not do	YouTube demo	paid version details
ClaudeAI	writing aide	search literature	AI Demo: Anthropic's Claude 3.5	$18/month for Pro; provides more searches and additional models
LitMaps	research; searching the literature	summarize literature; help with writing	Litmaps \| AI for Researchers	$10/month for Pro; provides unlimited articles and advances search tools
Co-Pilot	writing aide; chat bot	search literature (does provide sources, but they tend to be less-reliable consumer-level sources)	Get Started with Microsoft Copilot (Beginners Guide)	Provided free of charge for GW users

How AI works

Writing tips when using AI

Remember that AI is good at smaller, simpler tasks such as drafting and outlining and helping to find ideas. Don't expect it to be able to write for you (at least not well).
Be a Human in the Loop. Review AI's output the same way you would a study that you're peer-reviewing. AI may provide some decent ideas but the final word should be coming from the human writer.

General Pros and Cons of AI for research

Pros:

Good for brainstorming. Tools for drafting and outlining research papers, meeting agendas.
Can do simple tasks.
Some will provide sources and are transparent about where they draw their information from.

Cons:

Often inaccurate and requires effort on the part of the prompter/user to evaluate the output.
Not always transparent about sources/citations. Or will get a good citation, it may not be used in proper context.
Potential for copyright violations (pulling from work done by others).

Scholarly Publishing: Artificial Intelligence and Scholarly Communications