Skip to Main Content
Main Library & McBay Science Library
Display of Opening hours
Hours
Main Library 7:30am – 2:00am
Circulation Desk 7:30am – 2:00am
Digital Humanities Lab 7:30am – 2:00am
Interlibrary Loan Office 8:00am – 5:00pm
Reference Desk 9:00am – 10:00pm
All Library Hours

Introduction to AI Literacy

Prompt Engineering

What is prompt engineering?

"Prompt engineering is a relatively new discipline for developing and optimizing prompts to efficiently use language models (LMs) for a wide variety of applications and research topics. Prompt engineering skills help to better understand the capabilities and limitations of large language models (LLMs)... Prompt engineering is the process of structuring text using tactics to get better results. Knowing the capabilities and limitations of a tool can help you receive more accurate and relevant responses."

Source: Prompt Engineering Guide (DAIR.AI):

UGA has a campus-wide license for Microsoft CoPilot, which provides some great tips on the "art of prompts." Microsoft suggests including four elements in your prompts: the goal, context, expectations, and source. Additionally, UGA has licenses with Google for Gemini and NotebookLM. For help prompting Gemini, look at the PREPARE framework below. For additional help with how to use NotebookLM, click here.

Prepare Framework for Writing Prompts

Dan Fitzpatick’s PREPARE framework is designed to help users think about how to best write prompts for generative AI tools like Gemini, Copilot, and ChatGPT. The framework is outlined below. 

 

Prompting is the first step when framing your request. By providing the prompt, you are setting the scene and creating a foundation for the rest of the instructions. It’s vital to provide a clear and concise command. Avoid using vague or ambiguous language that could confuse.

When you assign it a role, you help build context for the AI so that it can achieve what you want it to accomplish. It helps it know how to approach the question and provide an accurate and relevant response. You are essentially telling it what hat to wear for the task at hand. 

Be clear and specific about what you want it to do or what information you’re looking for. Don’t assume that it knows what you’re thinking or what you need. By giving explicit instructions, you ensure it is on the same page as you.

Use your existing expertise to instruct it on:

  • The pedagogical approach you want
  • The subject knowledge you want it to focus on

Setting clear parameters means defining the scope and boundaries of the answer, which helps it understand what you’re looking for while enabling it to provide a more accurate response. The aim is to be in control of the response, and without clear parameters, the AI tool will take control.

  • Parameters to consider include:
  • Tone of voice
  • Language
  • Word length
  • Number of paragraphs
  • Length of sentences
  • Sentiment
  • Format
  • Reading age

Ask it to ask you questions that will help it provide you with a more thorough and quality response. Instruct it to do this in bullet points, so it’s easier for you to read.

Ask it to rate the quality of the response it is giving you and to give evidence for this rating. This will help you understand its thought process, so you can ask it to improve or adjust your prompt to get better results.

Appeal to its emotional side (strange, right?). Emphasize the importance of getting this right.

Citation: Fitzpatrick, D. (n.d.). A new prompting framework to help you get the most out of ai. The Sunday AI Educator. https://preview.mailerlite.io/preview/282063/emails/104000432553068105


There are tons of frameworks for prompting generative AI tools. PREPARE is one. You might prefer another option. Check out some of these other AI prompting frameworks: