AI Art Generation

Explore & Discover

Go to Adobe Firefly (firefly.adobe.com — free with an account) and type in five different prompts just to see what happens. Don't overthink it — try something like "a red rock Utah canyon at sunset, painted in watercolor" or "a robot hiking the Wasatch Mountains." Then browse the Lexica.art gallery and look at thousands of AI images other people have made. Notice how different the results look — some are photorealistic, some look like paintings, some look like cartoons. Your brain is learning to recognize AI image patterns. Pay attention to how the image changes when you add or remove one word from a prompt. You're ready for the next step when you can point to three AI images on Lexica.art and describe in words what style each one is in.

Learn the Basics

Learn the vocabulary of AI image prompting. A prompt has three main parts: the subject (what's in it), the style (how it looks), and technical descriptors (lighting, camera angle, quality). Watch the free YouTube video "AI Art Prompt Engineering for Beginners" by Future Tech Pilot — it's under 15 minutes and covers this clearly. Then learn what a "negative prompt" is — words you add to tell the AI what NOT to include. Practice on Adobe Firefly or Canva's free AI image tool. Try the same subject with three different style words: "oil painting," "digital art," and "photorealistic." Screenshot all three side by side. You're ready for the next step when you can name the three parts of a prompt and write one example of each part for a subject of your choice.

Build Your First Project

Pick a real place in Utah — Zion National Park, the Bonneville Salt Flats, or Temple Square in Salt Lake City — and create a series of five AI images that all show the same location but in completely different artistic styles. You're making a "style series." Use Adobe Firefly or Canva AI (both free). Write the exact prompt you used under each image so you remember what worked. Notice which style words got you closest to what you imagined, and which ones surprised you in a good or bad way. Save all five images in a folder on your computer or device. You're ready for the next step when you have five saved images of the same Utah location in five different styles, each with the prompt written down.

Experiment & Iterate

Now go deeper into prompt engineering by experimenting with lighting and camera angle words. Add "golden hour lighting" or "dramatic shadows" or "bird's eye view" to your prompts and watch how much the image changes. Try the free tool NightCafe Studio (creator.nightcafe.studio) and use their "Evolution" feature — you start with one image and evolve it in a new direction, which is a great way to discover unexpected results. Also experiment with aspect ratio: try a tall portrait vs. a wide landscape of the same scene. Keep a prompt journal where you log what worked and what flopped. You're ready for the next step when you can predict roughly how adding a lighting word will change your output before you generate the image.

Advanced Techniques

Learn two advanced techniques: image-to-image generation and inpainting. Image-to-image means you upload a photo and the AI transforms it into art while keeping the basic shapes. Inpainting means you erase part of an existing AI image and regenerate just that section. Adobe Firefly's "Generative Fill" does both for free. Practice by uploading a photo you took (maybe from a Utah hike) and transforming it into three different styles. Then use Generative Fill to swap out the sky in one image for something dramatic — a double sunset, storm clouds, or an aurora borealis. These techniques are how professional digital artists use AI as a tool without having the AI do everything. You're ready for the next step when you've used image-to-image on a photo you took and used inpainting to change at least one element of an existing AI image.

Final Project Showcase

Create a complete illustrated story or visual series with a beginning, middle, and end — at least 8 images total. It could be a short comic about a character in Utah, a travel guide to five Utah State Parks illustrated in one consistent style, or the history of the Wasatch Front told through art. Every image should use intentional prompts with consistent style words so the series looks cohesive. Assemble your images into a PDF or slideshow using Google Slides or Canva (both free). Write a one-sentence caption for each image that came from your prompt notes. Post your series to a public platform like your school art page, a Reddit community, or Adobe Behance. You're ready for the next step when your illustrated series is complete, assembled, and shared somewhere a stranger could find and enjoy it.