Prompting with AI
Unlike conducting a Google search, where you look for existing information through keyword searches, generative AI requires a more nuanced approach. When you use generative AI, you’re not just searching; you’re instructing the AI to create new content. This process necessitates providing detailed and specific information to guide the AI in producing the desired results effectively. This information is provided in the form of a prompt.
In this post we will explore the art of crafting effective prompts to enhance your interaction with AI tools.
What is a Prompt?
A prompt is how you instruct an AI tool to perform a task for you. This could include creating content, summarizing information, or conducting research. You can enter this prompt by typing or speaking directly into the interface of the AI tool. The tool then processes the prompt and generates a response based on the instructions you have provided.
You can think of a prompt as a recipe for an AI tool. Just as a chef uses a recipe to prepare a specific dish, the AI tool uses your prompt to produce the precise response you requested.
This analogy brings us to the importance of prompting.
Why is Prompting Important?
Effective prompting is crucial because it dictates the quality of the AI’s response. Think of it as instructing an intern or assistant: the clarity and detail you provide directly influence the outcomes you receive. By using precise language and providing the right context, you guide the AI to understand and fulfill your specific needs. To achieve a high-quality response, it’s essential to include key “ingredients” – the right words and structure – when phrasing your prompts.
Prompt Structure
To help you master the art of AI prompting, let’s break down a basic prompt structure. This template was adapted from the presentation titled “Unlocking the Power of Generative AI in Higher Ed,” presented at NERCOMP in Fall 2024:
Each element of the structure serves a specific purpose:
- Role: Defines the AI’s identity or the perspective from which it should respond.
Context: Provides background information that helps the AI understand the scenario or the specific details it needs to consider.
Task: Specifies what you are asking the AI to do.
Output: Describes how you want the AI to deliver the response.
You are a (role) ______________ (context) ______________. Please (task verb) ______________ as (output type) ______________.
The table below provides additional information on each element of the structure:
Role & Goal Who am I? | Context & Background What are the details? | Task What should I do? | Output How should I do it? |
Teacher | Do/Don’t | Create list | Bulleted list |
Professor | Examples | Suggest | Lesson plan |
Tutor | Demographics | Rewrite | Case study |
Editor | Audience | Explain | Essay prompt |
Example
Use the template to create a prompt of your own in 1-2 sentences.
Improve and Refine your Prompt
To enhance your interactions with AI tools, consider refining your prompts both during initial and follow-up engagements. Below is a “prompt refinement” prompt that not only helps improve your prompts but also your prompting skills. You can utilize this strategy from your very first interaction with the AI tool.
From now on, whenever I write a prompt, act as an expert prompt engineer and suggest a better version of that prompt to use. Then, ask me if I would like to use that better version instead.
This approach helps identify gaps in information, encourages reflection on your current prompt, and explores better phrasing options. By refining your prompts, you leverage the AI tool’s capabilities to deliver more precise and useful responses.
You can also consider attaching relevant documents or data in follow-up interactions. This additional context enables the AI tool to provide responses that are more specifically tailored to your needs.
Based on above, use the (attached resource) ______________ to adjust the response for ______________.
Example
Below you will find the UDL 3.0 guidelines referenced in the example above. Feel free to download this document and use it to improve and refine your original prompt.
Examples of Academic Prompts
- I have encountered problems engaging with the students in my college literature class. It appeared that most of them do the readings but don’t participate actively in the discussions. Can you give me ideas that will help me and the students have more engaging discussions?
- Pretend you’re a professor teaching an introductory literature course at a large university. Create five ways that you can engage with students. Include ideas for class activities, assignments, examinations, and interactions with both the large group and small groups of students.
- You are an experienced university-level teacher in xxxx topic. You are in charge of a class with xx students in their sophomore year . Try five creative ideas to break the ice, engage and motivate students to be more active in class based on todays topic (context). Use references from xxxxx (whatever is viral or trending online) to make it more fun.
- I am an instructor of a freshman-level introduction to business course. The course is three credit hours and most students have some general business knowledge. Help me create five engaging activities that reflect the learning objectives X, Y, and Z, include group and individual work, maintain active learning, and encourages input by me at the end.
- I am an academic librarian teaching a research skills course. Create five ways for me to engage with undergraduate students at a small, private liberal arts college. These are first year students and every student has a laptop and internet access. Provide detailed examples. Gamify these ideas to increase student engagement.
To continue exploring prompts, a great prompt repository can be found here with additional examples, academic and otherwise, for you to test.
Try these techniques in your next interaction with AI. We’re eager to hear about your experiences!