TL;DR: This post dives into the basics of AI prompting, explaining how K-12 educators can effectively communicate with AI tools to enhance classroom activities and lesson planning. By providing clear instructions, known as prompts, educators can guide AI systems to generate useful content and support student learning. The post also highlights the importance of context, role, audience, and format in crafting effective prompts and offers tips on how to practice and improve prompt engineering skills. Several prompt libraries are recommended for further exploration.
Are you wondering about all the hype and excitement around AI and its implications for education? Be sure to start your reading journey with my first post in this series, "What the Heck Is AI Anyway?". This post gives you a comprehensive explanation of what AI isand how it works. I would also recommend you check out the book, Co-Intelligence by Ethan Mollick. It is, hands-down, the most accessible, easy to understand explanation of AI and its potential influences on our work and lives. It is a very quick, enjoyable read!
If you have at least a basic understanding of AI, it's time to invest some time playing with it! While you're reading this post, go ahead and pull up a chatbot like ChatGPT, Perplexity, Google Gemini, or Claude Sonnet so that you can test out some of the strategies and prompts I'm going to share with you. If you want to really get crazy, pull up all of them and compare their responses. This will help you begin to choose your favorite chatbot for your different purposes, since each chatbot has been trained on a different dataset and may be better at some tasks than others.
So, what is prompting? To help me define it, I asked all of the chatbots above to help me out. This is what I asked: "Give a concise definition of what ai prompting is. Your audience is K-12 educators." Here is how each responded:
ChatGPT: "AI prompting is the process of giving clear and specific instructions to an AI tool to generate desired responses or perform tasks. In a K-12 setting, this means asking the AI questions or giving commands in a way that helps the AI provide useful information, lesson ideas, or student support efficiently and effectively."
Perplexity: "AI prompting refers to the practice of providing specific instructions or questions to artificial intelligence (AI) systems to elicit desired responses or generate content. In an educational context, this can involve guiding AI tools to assist with tasks such as:
Generating Ideas: Helping students brainstorm topics for essays or projects.
Providing Explanations: Offering clear explanations of complex concepts.
Creating Assessments: Assisting teachers in developing quizzes or lesson plans.
By using effective prompts, educators can harness AI to enhance learning experiences, foster creativity, and support individualized instruction." *It is worth noting that in addition to giving me the definition, Perplexity also provided a list of related prompts that I might want to explore related to the topic. These suggestion included how AI prompting can be used in a classroom setting, examples of effective AI prompts for students, tools/platforms that can help generate AI prompts for K-12, how AI prompting differs from traditional teaching methods, and potential benefits of using AI prompting in education.
Claude: "AI prompting is the process of crafting specific instructions or questions to guide an artificial intelligence system, like a chatbot or text generator, to produce desired responses or outputs. For K-12 educators, it's a skill that can help them effectively use AI tools to create educational content, generate ideas, or assist with various teaching tasks."
Gemini: "AI prompting is giving specific instructions to an AI system to get a desired result. It's like asking a very smart assistant to do something for you by telling them exactly what you want. For example, asking an AI to write a story about a friendly robot is a prompt."
Essentially, a prompt in AI is the information that you give it in order to get a response. Being an efficient prompt engineer is very important because it will save you time, ensure your are receiving the best responses, and to help reduce the carbon footprint left when we use AI (read more about that here ). When I heard about the effect of AI on energy usage, I may have freaked out a bit and contemplated becoming a pescetarian; I cannot give up AI at this point because it has become such an integral part of my workflows. But, I digress.
To become a good prompt engineer, there are a variety of frameworks that you can use to ensure efficiency and valuable AI interactions. This graphic, created by Shushant Lakhyani for Medium.com presents 10 different frameworks for prompting:
Click Image to Expand
If you look at all of these, you'll begin to notice that they all have something in common. From these commonalities, it is clear that a good prompt will have:
Plenty of context. Other than PII (Personally Identifiable Information), you can never really give AI too much information. Anything that you would want to know if you were helping someone with this topic, you should share with the AI.
AI's Role (Role Prompting): Who does the AI need to be for you? An expert, a specific person, someone within a specific field of work or knowledge, etc.
Audience: Who will receive or benefit from the AI's response?
Format: What do you want the output to look like? For example, do you need a lesson plan in a specific format, a table, a song, a bulleted list, a recipe, etc.
Taking a little extra time up front to craft a good prompt will save you time and frustration on the back end. When I am partnering with AI on projects, I will typically open up a Google Doc and draft my prompts there first, before copying and pasting them into the AI. I have found this helps me get all of my thoughts out first and then I can rearrange them into a cohesive prompt, using the order of information I shared above.
Something else to keep in mind when prompting is the complexity of the task your asking the AI to perform. If your task requires multiple steps, you'll want to break down the tasks into it's simplest steps, order those in sequence, and then prompt the AI one step at a time. This is called prompt chaining, and ensures efficiency in the output and reduces the amount of energy used to create the responses. You can see an example of prompt chaining I used when I was coding a Google Spreadsheet.
💡 Pro Tip: After creating your account for a chatbot, go into the settings and complete the set-up so that you don't have to repeat some of the same context information over and over again.
You may be wondering how you become better at prompt engineering. One way is to simply practice. A good rule to keep in mind is to never "take" the AI's first response. You should always carry out at least 3 to 5 interactions, requesting edits, additions, explanations, etc. The more you "converse" with the chatbot the more you'll begin to understand how it "thinks" and makes connections. Another good method would be to explore some of the prompt libraries that are available. I love looking through prompt libraries and noting down words and phrases that I may not have ever thought of using. Here are some prompt libraries for educators for you to explore:
💡 Pro Tip: Ask your favorite chatbot to generate the prompts for you!
Want to practice your prompting? See how close you can get an AI image to match another using the fun tool, . Leave your score in the comments below. My highest so far is 78%.
I would love to hear any prompt engineering tips and suggestions you might have! Please share them and any of your favorite AI tools in the comments below. Be on the lookout for the next posts in this AI series!
AI Series:
Would you like exclusive access to member's only content, freebies, and updates? Be sure to join the mailing list for my once per month newsletter!
AI Disclaimer: Blog post title and images created in collaboration with ChatGPT.
Comentarios