Skip to Main Content

Generative AI for Research

This Research Guide provides information on the use of Generative AI in academic papers and research, and provides guidance on the ethical use of Generative AI in an academic setting.

What is Generative AI?

Generative AI—often referred to as GenAI—is a form of artificial intelligence capable of producing new content based on patterns learned from large datasets. A wide range of Generative AI tools are being developed at a rapid pace, these web-based tools use algorithms, data, and statistical models to draw reasonable inferences to create content of its own.  This output can include text, audio, images, and even video.  They are not search engines but rather trained chatbots, they will always want to give you an answer - even if it is incorrect.

These tools are increasingly capable of generating sophisticated content and are expected to significantly influence fields such as business, journalism, the arts, and education. As with any emerging technology, experts are both optimistic about the innovative possibilities and cautious about the ethical implications surrounding its use.

For a brief introduction to how generative AI works, we recommend the short LinkedIn Learning video: [What is Generative AI?]

Make sure it is ethical to use AI (Check your syllabus for your faculty's AI Statement) and fact-check any content and sources you plan to use in the work you share with others or publish that has been generated by AI.

There are many different types of generative AI that can create text, images, sound, video, and more. This page describes common types of generative AI.

Text generators Icon showing two chat bubbles with text on a mobile phone

Text-based generative AI tools create new text that is similar to the data they were trained on. The training process for these AI chatbots involves consuming large amounts of text from data from webpages, books, and other sources, then analyzing the text to find patterns and relationships in human language. Because of this training process, these tools are commonly referred to as Large Language Models (LLMs). They use probability to predict which words should appear in sequence. As Stephen Wolfram explained, “it’s just saying things that ‘sound right’ based on what things ‘sounded like’ in its training material.”

AI chatbots can produce essays, blogs, scripts, news articles, reflective statements, and even poetry. 

Some chatbots rely on their training data to produce content, while others are grounded in a source of facts.

Image generators Icon showing a landscape picture

This type of AI learns through analyzing datasets of images with captions or text descriptions. If it knows what two different concepts are, like a cat and a skateboard, it can merge those concepts together when prompted to create an image of a cat on a skateboard.

Generative AI image tools can produce diverse images in a range of media, everything from photorealistic oil painting style to anime.

Sound and music generators Icon showing a computer menu with music playing

AI music generators analyze music tracks and metadata (artist name, album title, genre, year song was released, associated playlists) to identify patterns and features in particular music genres. They may also be trained on song lyrics. If a music generator has only been exposed to one type of music (e.g., classical), then the music it generates will sound similar to those works.

Video generators Computer icon with video playing

Creating a video typically requires the use of audio, visual, and text elements. Some generative AI video programs have harvested existing videos to learn how to create new ones, while others have sourced the three elements to create video from audio, visual, and text sources. There are even generative AI video programs that have been trained to use video editing software, so they can apply effects to a video that you have created, such as adding captions, transitions, and animations..

Research discovery and explanation generators Computer icon with a magnifying glass over text

Some generative AI tools can automate parts of the research process and make long, complex texts easier to decipher. This type of AI often analyzes research papers that users upload to extract key information or summarize a paper.

What is prompting?
Simply, it's what you type into the chat box.

The way you prompt makes a huge difference in the output that any generative AI platform will give you. 

Start by asking questions about what you know.

By asking question about a topic that you are familiar with, you will see what the generative AI model gets correct and where it hallucinates (lies).

 

Always verify the information it gives you.

Think of generative AI as your personal intern. They need very specific instructions, and they need you to verify the information.

 

Generative AI sometimes makes things up. 

That's because it's designed to write in a way that sounds like human writing. It's not designed to know facts.

 - Use a Research AI tool to help analyze academic papers.

 

Tips for writing effective prompts

  1. Give it some context or a role to play.
  2. Give it very detailed instructions, including how you would like the results formatted.
  3. Keep conversing and asking for changes. Ask it to revise the answer in various ways.
  4. The longer you work with a Generative AI tool the more like you the generative text will become.

Leo Lo, a librarian and professor at the University of New Mexico, has developed another popular strategy for prompt engineering known as the CLEAR Framework1. It encourages users of generative AI to construct prompts that are Concise, Logical, Explicit, Adaptive, and Reflective.

  • Concise - Make sure the prompt is centered on a specific task and doesn't include unnecessary information.
  • Logical - The information provided to the AI tool should be given in a logical order.
  • Explicit - Be specific and avoid ambiguity.
  • Adaptive - Adjust the prompt if you don't get the response you need after submitting it the first time.
  • Reflective - Take a critical look at the prompts you've constructed and the type of information you received in return. What do you notice about the way a particular generative AI tool responds to different types of prompts? Consider how you can use this information in the future.

1Lo, L. S. (2023). The clear path: A framework for enhancing information literacy through prompt engineering. The Journal of Academic Librarianship, 49(4), 102720. https://doi.org/10.1016/j.acalib.2023.102720

Fact-checking is always needed

AI "hallucination"
The official term in the field of AI is "hallucination." This refers to the fact that it sometimes "makes stuff up." It is important to know that the information generated by AI wants to be correct, that doesn't mean that it is always factual.
 

Web search results as grounding
If you want information to be more accurate, feed your AI a trusted website and there will be less hallucinations.  That's because it can search the web, read the pages it finds, and use the AI model to summarize those pages, with links to the pages. It may sometimes make a mistake in the summary, so it's always good to follow the links to the web results it found.

 

Use An AI Research Tool to help conduct research

Use the AI Tools for Reserach Analysis tab to find the right research helper for your assignment.