Contact: kevin.sebastian@unlv.edu
This guide was developed in part by using ChatGPT-4o, a generative AI language model. It assisted in drafting content, providing recommendations, and refining language to ensure clarity and coherence. All AI-generated text was reviewed and edited for accuracy and relevance to the topic.
Try out Harvard MetaLab's Large Language Model (LLM) tutorial. The best way to learn about the capabilities and the limits of generative AI is by directly engaging with it. Follow the tutorial to help you get comfortable with and prepare to knowledgeably utilize generative AI.
Prompt engineering is the process of crafting specific inputs or “prompts” to guide generative AI models like ChatGPT to produce desired outputs. Think of it as a way of “communicating” with AI: the more precise and thoughtful your instructions, the more relevant and accurate the AI’s response will be.
For university students and faculty, prompt engineering can be compared to framing a research question or setting a problem statement. Just like how a well-defined research question leads to more focused and meaningful research, a well-constructed prompt helps the AI generate responses that are more aligned with your needs. For example, if you’re asking the AI to summarize a complex article, including details like the key themes or specific sections you’re interested in will lead to a more relevant summary.
Effective prompt engineering involves understanding the model’s capabilities and limitations, being clear and concise in your instructions, and iteratively refining your prompts based on the responses you receive. This skill can be particularly useful in academic settings for tasks such as generating study materials, exploring creative ideas, or even drafting initial outlines for research projects.
What follows is a "prompt anatomy" developed by Google for their Skillshop course "Generative AI for Educators." (The explanations and examples have been modified from the original Google Skillshop course.)
When optimizing prompts for generative AI, think about using five prompt-writing building blocks. Remember them by thinking of the “PARTS” of a prompt:
Explore each part below to learn more.