Generate visual metaphors to support learning

One of the ways I use images in learning experiences is with visual metaphors. When I develop training on abstract concepts, concrete images often do not make sense. Symbolic images can be more effective than literal images in these situations. However, stock images tend to reuse the same symbols and visual metaphors still and again until they become tired and less effective. You know those: images of hands in different colors combined in a circle for diversity, someone climbing a mountain to overcome adversity, a bulb on a person with a high finger to show an idea, etc.

One way I use AI to improve the quality of what I produce is to generate unique visual metaphors to represent abstract concepts. These images can support concepts and perhaps even make people establish different mental links with content.

Replacement of an image stock by a visual metaphor generated by AI-Ai

A visual metaphor that I used on my blog linked to the concept of character challenges and to overcome adversity.

In my previous blog header, I used an image in stock of someone climbing a large staircase. At least, it was not an image of climbing a mountain or jumping over a gap between two large rocks! It's still quite boring. It very obviously looks like a photo.

When I recently updated this blog article on Character challengesI also updated the header image. This time, I generated an image in Media With a flower pushing from a desert landscape. It is always a recognizable metaphor, but it is not as common. In addition, this is a unique image, generated just for this blog post. It does not look like all other stock images.

Showing a process with visual metaphors

For one of my projects last year, I created a short activity linked to a problem solving process. The activity was aimed at primary school students, so I wanted colorful images that would resonate with a younger audience.

The 4 -step process (paraphrasus and adapted to the real content of my client):

1. Identify the problem.
2. Think of your options and what you could do.
3. Choose one of the options.
4. Think and think about what happened.

I had the idea of ​​doing something new with this exaggerated visual metaphor to jump over a cliff between two mountains. What if it was a small figure or a doll that jumped between a gap between two high towers of blocks?

What I finished did not quite correspond to my mental image, but I have always loved the outing. In this case, the image to “identify the problem” is a small figure on the ground between a space of two rounds of blocks.

Then, I generated an image of several objects that could be used to fill this gap: more blocks, string, a chain, etc. These are my metaphor for brainstorming possible solutions.

In my metaphor for choosing a solution, I showed a child's hand holding a red wooden block. (This image met my needs at the time, and it worked well on a smaller scale in the context of the learning activity. But in the large -scale image, you can see the faults of physics where the block is on the hand. Given the improvements of the IA image generation in the past year, I could probably create something better now.)

Finally, I showed the tiny wooden figure above a bridge between the two block towers as a visual metaphor to think about the solution.

Although the images have faults, I was able to use Midjourney to generate this series of images for a unique visual metaphor. These images have a similar appearance with wooden cubes with rainbow combinations and the depth of shallow field.

I could have spent hours looking for stock image libraries for a series of images to represent this very specific process. And frankly, I am not sure I could find four images in a similar style for these specific ideas, no matter how long I was looking for. I should probably have been satisfied with one or two photos showing part of the process.

This is an example of the reason why I use the generation of AI images: it allows me to do things that were not possible before. It's not just about saving time. (Honestly, sometimes I spend a lot of time and multiple attempts to generate images.) The generation of AI images also allows me to improve the quality of what I create with specific best quality images.

Generate visual metaphors ideas

I sometimes try simply to provoke a visual metaphor without specifying what I want. For example, these are some results for a “visual metaphor for problem solving”. As a rule, I find that I cannot use any of these images as they are, but sometimes they help to arouse an idea that I can refine with other prompts.

I use LLM like Chatgpt and Claude to help me think about ideas for these visuals.

What are some creative visual metaphors for overcoming challenges? I don't want to do the same stereotypical mountain climbing. Give me a minimum of 5 different options.

Here is the list of metaphor ideas generated by Claude.

1. Disentangled helmet

2. A seed that breaks through concrete

3. Restore an old paint

4. Paper Boat Navigating rain gutters

5. Solve a Rubik cube

6. Kintsugi glass (Japanese art to repair broken pottery with gold)

For brainstorming with AI, I often use it to generate ideas like this, even if I do not use any of these ideas exactly. It helps me to arouse other ideas or to establish links that I had not considered.

I use similar prompts to think about ideas for icons and other symbolic illustrations. Sometimes it is so difficult to find a way to visually represent a concept in a simple way, and AI helps this.

I also find that the generation of these symbolic images takes more time and iteration than generating simpler concrete images. I often need several attempts to refine my prompt. Sometimes I end up going to a different concept entirely because I can't really generate what I want (or I can't understand how to describe it so that the tool can understand it).

What challenges are you confronted with the IA images generation?

If you tried to generate IA images, what challenges have you met? What questions would you like me to answer in future blog articles?

The idea of ​​this post came from a question that someone asked me to represent abstract concepts. I would also like to hear you talk about your questions and challenges! Let me know in the comments.

Related