Using Generative AI (GenAI) for Research

What is prompting?

Simply, it's what you type into the chat box. 

The way you craft your prompt, also called prompt engineering, makes a huge difference in the output that GenAI tools produce. These tips will help you craft prompts that generate the results you’re looking for.

Tips for writing effective prompts

1. Give the tool some context or a role to play.

2. Give the tool very detailed instructions, including how you would like the results formatted.

3. Keep conversing and asking for changes. Ask the tool to revise the answer in various ways.

Examples

1. A role could be, "Act as an expert in [fill in the blank]."
   Act as an expert community organizer.
   Act as a high school biology teacher.
   Act as a comedian.

2. Example prompt:

Act as an art historian. I’m writing a research paper for my History of Asian Art class, and I need help coming up with a topic. I’m interested in gender representation in Asian art. Provide background information on gender representation in Asian art history.

1.  Example of changes: (keep conversing until you get something useful)
   I found your background information on gender and colonialism useful. Provide a list of keywords and search phrases I can use to find relevant academic resources on this topic in library databases and Google Scholar.
   
   Or...

That background information was too broad. Provide more focused background information on the representation of women in Korean art during the Japanese colonial period.

In the following video, Dr. Daniel Anderson, Director of the Writing Program at the University of North Carolina at Chapel Hill, shares an approach to prompting GenAI that may help you think more deeply about your research topic.

Written content is adapted from “Student Guide to ChatGPT” by University of Arizona Libraries, licensed under a Creative Commons Attribution 4.0 International License. 

“AI Prompting and Thinking” by Daniel Anderson is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.

What is lateral reading? 

Lateral reading is a simple but effective method for evaluating any online content, including content produced by GenAI applications. Lateral reading is a fact-checking strategy where you open multiple browser tabs to verify information using trusted sources. 

Verifying information generated by a GenAI chatbot requires extra care because they:

• Generate responses based on prompts, so any misinformation or bias in the prompt can impact the output
• Often decontextualize information, making it difficult to trace its origins
• May pull information from both accurate and unreliable sources

Examples

The following videos from University Libraries at the University of Maryland demonstrate how to fact-check text, links, and scholarly sources using lateral reading.

What is the ROBOT Test? 

Each generative AI tool on the market is unique, so it's important to think critically when choosing one for your academic research. The ROBOT Test is a set of questions designed to help evaluate these tools:

• Reliability: How reliable is the information available about the AI technology? What is the goal of sharing information about the tool?
• Objective: What is the tool meant to help you accomplish? How does the tool's creator benefit from your use of the tool? 
• Bias: What types of bias might be present in the information the tool provides? Are there ethical issues associated with this technology? Does the tool's creator acknowledge bias and/or ethical issues?
• Owner: Who developed this technology? Who owns it? Who has access to it? Does everyone have equal access? What happens to information you enter into the tool?
• Type: What type of AI technology does the tool rely on? How does the tool work? 

Example

Librarians at Dartmouth College documented their process for evaluating ChatGPT 3.5 with the ROBOT Test. 

Written content is adapted from “The ROBOT Test” by Sandy Hervieux and Amanda Wheatley, licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

Evaluating GenAI output with lateral reading 

Content adapted from University of California Irvine Libraries’ Generative AI and Information Literacy LibGuide written by Stacy Brinkman and April Urban. 

UMD Libraries. (2023, August 17). AI fact checking scholarly sources [Video]. YouTube. https://www.youtube.com/watch?v=3nB4nRHdgUc 

UMD Libraries. (2023, August 17). AI fact checking text & links [Video]. YouTube. https://www.youtube.com/watch?v=XpbEbfMDsBk 

Evaluating GenAI tools 

Hervieux, S. & Wheatley, A. (2020). The ROBOT test. The LibrAIry. https://thelibrairy.wordpress.com/2020/03/11/the-robot-test 

Research Guides: Generative AI & Research: Example - Evaluate a Tool (ChatGPT). (2024, October 7). https://researchguides.dartmouth.edu/GenAI/evaluatetool