Skip to content

If you are exploring Artificial Intelligence for the first time it can all be a little overwhelming, especially figuring out how to prompt the machine and prompt engineering.

What even is prompting or prompt engineering?

Prompt Engineering is a way to make your results with artificial intelligence more reliable and useful. Large language models (LLM) are adjusting, and learning over time so understanding how to prompt them will give you better results more often.  

Sure, you can have a simple prompt; “make a quiz,” and you will get a result, but if you want a better result, there are principles to follow.  

Before we talk about those principles let’s take note that a prompt engineer is a current career and a career that our students will have as an option moving forward. Understanding prompting is one of the skills students will need to succeed in the future, whether to become a prompt engineer or in their everyday jobs.  Prompt engineers create prompts and help train the LLM to produce the best possible outcome.  According to Forbes, these jobs can earn a salary of up to $300K a year.

So, how do you make the BEST prompt?

B- Brief the Bot

The first part is easy; give the machine a “brief” or instructions on what you want it to do. Think about the outcome you want and ask the bot to do it.

Here is what this could look like: 

  • Please create a quiz about photosynthesis. 
  • Please create an itinerary for a field trip to Boston in the spring.
  • Please read the following text and answer questions about it when I ask.

E – Expertise

Next, inform the robot what it is an expert in or what role you want them to play. This gives the machine more context.  You can give the bot more than one role or more than one expertise.

Here is what this could look like: 

  • You are an expert in NGSS (Next Generation Science Standards) for middle school.
  • You are a kind and caring teacher.
  • You are Katniss Everdeen from The Hunger Games.

S – Specifications

Now that you have a brief and expertise the next thing you want to do is give the machine specifications on how you want the appearance and details of your results.  You have to think about what and how you want your results.  You can include tone or structure in your prompt.  

Here is what this could look like: 

  • Put the response in a table with two columns.
  • Write this in a formal and caring tone.
  • Write this in 100 words or less.
  • Make the complex ideas easy for a *14-year-old to understand.

*When you are giving a LLM instruction for a grade level make sure you use the age.  This will give you better results than a grade level because it’s more specific. 

T – Thanks

Many ask why I say please and thank you in my prompts. I joke and say “It’s so when the robots take over they remember that I was the polite one,” In all seriousness, some studies show that using “emotional stimuli,” can produce better results.  This could be telling the bot why this outcome is important to you and encouraging the bot.

Here is what this could look like:

  • This must be right to help teach my students about photosynthesis 
  • It’s crucial that my students understand the point of view of this character. 
  • Make the tone happy and upbeat.
  • Are you sure that’s your final answer? Believe in your abilities and strive for excellence. Your hard work will yield remarkable results (This one comes from this study).
  • *Please…..
  • *“Prompt” Thank you

*This study focuses on emotional stimuli in prompts like the first four examples, but adding please and/or thank you to your prompts can influence the tone and style of interaction. Think about when you use please and thank you with humans and how that changes an interaction and tone.

Keep the HUMAN in the loop

After prompting the machine you must keep the HUMAN in the loop!

Once you’ve hit enter and get your output don’t forget to keep the human in the loop and make it yours.  

We’ve created this acronym to help you remember: HUMAN

H – HALT – Is this something you want to use AI for? Is your outcome appropriate for the task? Is it an ethical use? 

U – UNDERSTAND – Understand and check for bias.

M – MAKE IT YOURS! – Does this sound like you? What can you add to this or do to it to make it yours?

A – ACCURACY – Check to make sure it’s accurate. Do you have evidence/research to back up the information produced?

N – NEXT – What will you do with it next, do you need to disclose that you used AI?  Do you need to cite it? 

Kate goes into more detail about keeping the human in the loop with her students and the ethical use of of AI in this post.  

Prompt Engineering FREEBIE

Finally the FREEBIE! We want to help you learn prompt engineering and help you teach your students about it.

Kate shared this in her last post and I want to share it again to include both keeping the HUMAN in the loop and to make the BEST prompts.

AI Prompting Posters
3 Different Prompting Poster Versions Based on Printing Needs

Lastly, keep practicing. Practice makes perfect. The more you prompt the machine the better you will get at prompt engineering, and the more you’ll discover how it works.

-Stay Curious, Stay Innovative!

Leave a Reply

Your email address will not be published. Required fields are marked *