Beyond Chat: Discovering the Power of Image-Based Prompts in LLMs


I have been having so much fun exploring AI tools and learning how they can help teachers save time and achieve learning goals. So, I have seen a lot, but one thing I came across recently really blew my mind. 

Ever since I learned about Large Language Models, I’ve pretty consistently used ChatGPT. However, at the Future of Educational Technology Conference this January in Orlando, I attended a session that showed some of the functions of other LLMs. Gemini (formally known as Bard) allows you to upload an image to be a part of your prompt. 

I decided to test out this functionality, and I was not disappointed. I love making hyperdocs based on the 5E model. I took my favorite template – which you can check out here – took a screenshot and uploaded it to Gemini. Usually, I put more effort into my prompt engineering, but just for the sake of my little experiment, I entered this prompt: 

“I would like to make a 5E model lesson for my 11th grade honors environmental science class. The topic is climate and sustainability. Please create a lesson for them that follows the guidelines in an image.”

And much to my amazement, it scanned the image and planned out a 5E Model lesson that followed the guidelines in the image! Maybe I am late to the party, but I was so surprised and impressed. Now my mind is racing with all the possibilities and extensions. Stay turned! 

You can check out the output for yourself here

Most Recent

Leave a Reply

Your email address will not be published.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