Like so many things in life, what you get from AI image generators is only as good as what you put into them.
In the world of generative AI, these instructions we give AI tools are called prompts – and the types of prompts you use directly impact the images you get in return. Different AI image prompt modifiers can help you create the perfect images for your use case.
In this article, you’ll find 10 different types of AI image generator prompt modifiers, each with several different examples of terms you can use to produce more interesting, engaging illustrations and photo-like images for any purpose.
I used Craiyon for the image examples throughout this post. It produces nine variations for each prompt, allowing me to create unlimited watermarked images with the free version.
Whether you’re generating images for marketing materials (blog header images, advertisements, social content, etc.), using AI to illustrate a book you’re writing (as a friend of mine is doing right now!), or just experimenting with image generators for your enjoyment and amusement, these prompts will help you produce more varied output.
1. Subject Matter
Your subject is what’s in the image – the focus or topic of the picture. Try a combination of two or three descriptors for your subject to give the AI a better idea of what you’re looking for. It could be a thing + a specific place, a type of animal + an action, or an object + a location, for example.
Some other examples of subject matter could include:
- a person or specific body part, eg.: face, feet, eyes
- plants and animals
- elements in nature: the moon, stars, ocean, beach, trees
- intangible concepts such as beauty, wonder, awe
- more common concepts such as family, community, together
- buildings: house, skyscraper, hospital, office
- inanimate objects: a vase, fruit, chairs, a city park bench
Now, let’s add some more modifiers to change our image output.
What is your subject matter doing? They could be reading a book, taking a class, flying an airplane, picking up trash. The subject matter doesn’t have to be one person or thing but could be a flock of sheep or a group of people. So what are they doing?
Think of all of the action verbs you learned in school: fall, jump, run, crawl, eat, drink, hit, punch, swing, go, stop, walk, sleep, wash, open, close, push, pull, play, hug, tickle, dance, shake, sit, stand, toss, catch, etc.
Give the AI a specific action for your subject matter to take and see what it comes up with. I made mine a baby (subject matter) dancing (action):
3. Photographic style
Changing the photographic style can result in an entirely different image.
Try adding any of these to your image generator prompt:
- still life
- macro photography
- composite, etc.
Here’s an example of two images created using the same subject matter, plus the photographic style prompt modifier “macro photography” for one and “landscape” for the other:
Changing the colour in your prompt can result in some fun and wildly different results, especially if you use an unexpected colour variation for the subject matter:
Lighting can completely change the mood and tone of your image. Try adding a lighting modifier to your prompt:
- golden hour
- dawn or dusk
- flat lighting
- overhead lighting
- butterfly lighting
- rim light
- loop light
- ring light
- flash bulb, etc.
Look at the difference in these images achieved by adding “butterfly lighting,” “overhead lighting,” and “backlit” to the same prompt:
6. Artistic style
This is another fun image prompt modifier that can dramatically change your output. Some examples of artistic style include:
- art deco
- art nouveau
- film noir
Here’s what Craiyon thought two different birthday parties might look like, one in steampunk and the other in art deco style (interestingly, all subjects in both sets of nine variations were women even though we didn’t specify gender):
7. Material or medium
Medium refers to the substance used to create a work of art, while material can refer to what your subject matter is made of. Using one or both of these can result in completely different outputs.
Try using these modifiers in your generative AI image prompts and see what you come up with:
- oil paint
- pencil sketch
- acrylic paint
Check out these examples using pencil sketch and oil paint as the medium:
And these using ceramic and yarn as the material the subject (our iguana) is made out of:
Perspective refers to how the subject is presented in its space. Here are a few different perspectives you can use in your generative AI prompts:
- aerial or atmospheric
- bird’s eye view
- diminishing scale
What can adding a perspective prompt do for your AI images? Check it out – here’s what Craiyon produced as examples of a busy highway from an aerial perspective:
And we get a completely different output with the same subject matter but changing the perspective to diminishing scale:
The mood of an image can completely change how the subject matter is represented.
Here’s what this AI image modifier looks like in action:
10. Era or Historic Period
What time period is your image set in? Some examples of historic periods and eras you may want to experiment with include:
- Victorian age
- ancient Greek or Roman
- Middle Ages
- digital age
- Edwardian era
- atomic age
You can create some fun variations by placing subject matter in a time and place it doesn’t belong; for example:
Bringing It All Together with Unique, Descriptive AI Image Prompt Modifiers
Now, combine several of the above modifier to see what you can create in AI.
Try changing one modifier at a time, or introducing new ones and taking some away, until you get what you want.
Avoid being too descriptive, as you want to allow the AI to exercise a bit of its own “creativity.” Remember, it’s just tapping into its knowledge base – the data it was trained on or has access to – and generating images based on its perception and understanding of your prompts.