How do you create images and designs with AI?

Start creating images without complexity

Start with a simple description

Models are now very capable. You can open ChatGPT or Gemini, write a simple description of the image in your mind, and get a professional image or design without Photoshop experience or an expensive camera.

For example, say, “Create a cartoon image of children in astronaut suits on a planet named after Al-Zohour School,” and it will get started with you. You do not have to memorize the components of a long prompt or use lots of technical terms on your first try.

If you like the first result, great. If you want more control or the same quality repeated across a set of images, that is when you start adding details:

Who is the image for?
What will you use it for?
What is the most important element that must appear?
Do you want it realistic or illustrated?
Where will you publish it, or what will you print it on?

If you are a teacher, for example, you could start with a simple prompt about space. After seeing the result, explain that you want a cover for a school activity for ten-year-olds, showing children in spacesuits in a simple cartoon style, with clear space for the school name and activity title.

The image has a purpose

You are doing more than describing something attractive. You are defining who the image is for and how you will use it.

Start creating images without complexity

3 ways to create an image

Before choosing the model or writing the prompt, define what you want to do. In the lecture, we used three main approaches.

1. Text to image

You have an idea, turn it into a written description, and the model builds the image from scratch. This works when you have an idea in mind but no existing image that expresses it.

You can use it for a cover, educational scene, ad, imaginary character, thumbnail, or illustration.

Try this prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Create a cartoon cover for a school activity titled “Al-Zohour School Planet.” Show five children aged 9 to 11 in the center, wearing colorful spacesuits and standing on a friendly-looking planet, with a simple space background and cheerful colors. Leave clear space at the top for the title. Aspect ratio: 16:9.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

2. Image to image

Here, you are not starting with a blank canvas. You upload an image, then explain its role to the model.

Style reference: You like the illustration style, element arrangement, or colors.
An element that must appear: A product, character, logo, or personal photo.
An image needing an edit: You want to change the background, remove an element, adjust the dimensions, or add text.
Multiple references: An image of the character, an image of the product, and a third image for the desired style.

Most importantly, do not assume the model will understand each image’s role on its own. Say clearly: “Use the first image for the character, the second for the product, and the third as a color and style reference.”

3. Image to text

This reverses the process. Instead of writing a prompt to get an image, upload an image you like and ask the model to analyze it and write a description that helps you create a similar result.

Try this prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Analyze the image as a visual reference and write a detailed description of its illustration style, colors, lighting, composition, camera angle, and font. After the analysis, turn the description into a prompt I can use to create a different idea with the same feel, without copying the original elements.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

The goal is not to copy someone’s work. It is to understand what you liked about the image and turn it into clear directions for an idea of your own.

Understand models and websites

The same model may be available on several websites

Let’s distinguish the image model from the website you use it on.

For example, there is ChatGPT’s image model and Google’s Nano Banana models. The same model may be available on its official website and on another website or tool.

You can use Nano Banana models through Gemini and through Google Flow. Flow is not a different image model from Gemini here. It is another Google workspace providing the same family of image models in a format suited to projects, images, and videos.

You can use ChatGPT’s image model through ChatGPT, and you may find other websites offering it.

Do not let website names distract you. Focus on which model you are using and whether its results suit you.

Start with the website you find easiest. Write your description, see the result, and edit it through conversation. If you want a project with many images or plan to continue into video, you could work in Flow. If you are comfortable in ChatGPT or Gemini, keep working there.

Interfaces, free limits, and credits change over time, but the basic approach stays the same: write, try, edit, and download the result suited to your use.

Get more control over the result

A better result you can repeat

A simple description is enough to start, and today’s models can produce excellent results from it. The next section is not required for your first image. Use it when you want a more precise result, a consistent character, or a set of designs with consistent quality and style.

You do not need to memorize a long formula as though it were a rule. Whenever you need more control, answer the questions that affect your result.

1. What type of result?

Say from the start: a realistic photo, illustration, poster, infographic, cover, thumbnail, carousel, or educational design.

2. Who or what is the main subject?

Describe the character, product, or main idea. For a character: age, clothing, expression, and action. For a product: shape, color, material, and details that must not change.

3. What is happening, and where?

The scene is more than a character standing in empty space. Explain the action, location, background, and important surrounding elements.

4. What is the style?

Realistic, cartoon, Clay, Watercolor, 3D, Editorial, or Minimal. If describing it is difficult, upload a reference and specify exactly what you like about it.

5. Lighting and colors

Define the overall mood: cheerful colors for children, calm professional colors, morning light, cinematic lighting, or brand colors.

6. Dimensions and use

Say where the image will go: a Story, post, YouTube, display screen, printed sheet, or cover. If you know the ratio, write it: 1:1, 4:5, 9:16, or 16:9.

7. Text inside the design

If you want text inside the image, write it exactly in quotation marks and specify its location and importance. If the content is long, approve the text first, then request the design. Image quality does not guarantee that every piece of information is written correctly.

A practical template

Use and adapt the template

Copy the prompt as it is, or tailor it with your details first.

Tailor the prompt to your workOptional: add your details and the prompt adjusts itself

Image typeAudience and useCharacter or product descriptionActionLocationStyleColorsLightingSpace locationTextPercentageUnwanted elements

View full prompt

Create [image type] for [audience and use]. The main subject is [character or product description], shown [action] in [location]. Use [style], [colors], and [lighting]. Make the main subject the focus, and leave space [space location] to add the text “[text].” The design should be in Arabic and read from right to left. Aspect ratio: [ratio]. Avoid [unwanted elements].

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

You do not need every field every time. Use what actually affects the image.

Get more control over the result

A reference saves a long description

Sometimes you see a design on Pinterest and like its font, layout, or character style. However much you describe it, the model may produce something far from what you have in mind. A reference image shortens the process.

But you need to say what you want it to take from the reference. If you upload an infographic about tea and only say “Make one about milk,” the model might imitate elements you do not need or change important information.

Try this prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Use the uploaded image only as a reference for the illustration style, background colors, heading style, and section layout. Create a new infographic about the benefits of milk for children aged 10. Do not copy the original text or illustrations. Arrange the content from right to left and use short, clear phrases.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

Visual inspiration is not copying

Whenever you see a design you like, ask yourself: Was it the colors, font, spacing, icon style, composition, or how the information was presented?

Once you identify the element, you can use it as a reference without turning your work into a copy of the original design.

If the design includes another organization’s logo, watermark, or brand identity, do not ask to remove it so you can use the same design as your own. Use it as a reference and create a new design with your brand identity and content.

Adjust dimensions, Arabic text, and edits

Choose the dimensions from the start

A square image, Story, YouTube thumbnail, and printed poster are different things. If you leave the dimensions to the model and later try cropping, you might cut off the main subject or lose the space where you planned to add text.

Use	Common aspect ratio
Square post	1:1
Portrait post	4:5
Story or Reel Cover	9:16
YouTube Thumbnail or display screen	16:9

These are common ratios, not fixed rules. If the platform or print shop requests specific dimensions, use those.

Think about space before generating, too. If you will add a title on the right, ask the model to leave the right side calm and relatively empty. That space is part of the design, not a flaw.

Adjust dimensions, Arabic text, and edits

Arabic needs extra review

Image models have improved at writing text, but a letter can still change, a word can repeat, or the design can end up running from left to right.

Write the final text yourself before requesting the image.
Explicitly ask for a right-to-left design.
Reduce the amount of text inside the image.
Check every letter, number, and name before publishing.
If the text is central or long, use the model to create the background and elements, then add the text in a design tool that gives you finer control.

For an educational infographic, check more than spelling. Check the information itself, because the model may create an attractive design with missing information or a misplaced heading.

Adjust dimensions, Arabic text, and edits

Edit the result instead of starting from scratch

AI does more than create new images. You can upload an existing image and ask it to edit it as needed.

You can improve image quality and clarity, adjust lighting and colors, reduce noise, remove an unwanted person or element, change the background, or restore an old image with scratches and unclear details.

Try this prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Improve the image quality and clarity, adjust the lighting and colors naturally, reduce noise, and preserve people’s facial features and the original image details without exaggerating or changing their appearance.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

Start with a simple prompt, see the result, then add the details the model missed. If the image matters, have it make the edit in a new copy so you keep the original.

The first image does not have to be final. Treat it as a visual draft. If it is close to what you need, ask for one clear edit at a time:

Change only the background and keep the character unchanged.
Expand the space at the top of the image for the title.
Keep all elements and change the aspect ratio to 4:5.
Correct the written sentence.
Remove small distracting elements.

When should you start again?

Start again if the core idea is wrong: the camera angle, scene composition, or style choice. But if the problem is an extra element, text, or color, editing is often faster.

After each version, check faces, hands, text, the logo, product, and anything that must stay the same.

Turn the image into something you can use

From one image to a full set of content

The uses are almost limitless. Looking at your work, studies, or life will keep giving you new ideas.

For teachers

You can turn a difficult concept into a simple drawing, create an activity cover, review card, lesson character, an infographic summarizing a lecture, student certificates, or stickers for encouragement and follow-up.

You can also turn a concept into a complete illustrated story: establish the characters first, choose the style, then turn the story into connected images.

Try this prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Help me turn the water cycle in nature into an illustrated story of six scenes for a child aged 9. Suggest the characters and story sequence first. After I approve, write each scene’s description while keeping the same characters and cartoon style across all images.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

For content creators

You can create a consistent identity for a post series, a quote design, video cover, or carousel explaining an idea across connected cards. If you publish on different platforms, make a version suited to each size and reading pattern instead of posting the same design everywhere.

For business or product owners

You can create product images in several scenes, but you must preserve its real details. Prepare clear photos and tell the model what cannot change: packaging shape, color, logo, embroidery, or material.

A good image is also the first step toward a good video. With a clear character or product image and an organized reference, you can later use them in a storyboard or video scenes, keeping details more consistent.

For personal use

You can design a phrase you like, an invitation, a card, a keepsake image, or an illustration of an idea you have in mind.

You can also upload your photo and ask to change the clothes or style. Keep privacy in mind, and do not use anyone's photo for something they have not agreed to.

Turn the image into something you can use

Create an infographic from accurate content

An infographic is more than an image with words on it. It is a way to condense information and make its structure clear at a glance.

Define the title.
Choose three to five main ideas.
Shorten each idea to one or two lines.
Check the information.
Define the audience and dimensions.
Then ask for the design.

A ready-to-use infographic prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Design an Arabic infographic titled “How do you create a professional image with AI?” for beginners. Divide the design into six short sections: image creation methods, choosing a tool, describing the main subject, using a reference, choosing dimensions, and reviewing the result. Use simple icons and navy blue and green colors, and arrange the content from right to left. Do not add information yourself; use only the text I sent. Dimensions: 4:5.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

Ask the model for the design, but review the final content yourself.

Turn the image into something you can use

A carousel is one idea divided well

A carousel is a series of cards, but that does not mean dividing a long paragraph across five images. Every card needs a clear role.

Cover: How do you create a professional image with AI?
Text to image: Write your idea and define the audience and use.
Image to image: Use a reference for the style or an element that must appear.
Image to text: Analyze the design and understand its components.
A clear prompt: Character, location, movement, style, colors, and dimensions.
Before publishing: Check the Arabic, information, and details.

A ready-to-use carousel prompt

This is a ready-to-use example. Copy it as is, or adapt the text to your work.

View full prompt

Make an Arabic carousel of six cards using the attached content. Keep the same visual identity, character, and colors across all cards. Make each card clear on its own, with simple visual continuity between cards. Arrange the text from right to left and do not change or add information. Dimensions: 4:5 for each card. Show each card's content for review before creating the images.

If you are missing information that would change the result, ask me first.
Before you deliver: check the Arabic text, dimensions, and details I asked you to preserve.

Create and practice yourself

A simple workflow for any image

Define the goal: Who is the image for, and where will you use it?
Collect the content: Correct text, information, and data.
Choose the method: Text to image, image to image, or image-to-text analysis.
Prepare references: Style, character, product, or colors.
Write the prompt: Focus on the details that actually change the result.
Choose dimensions: Before generating the image.
Generate several attempts: Compare instead of sticking with the first result.
Edit: Fix one clear element at a time.
Review: Arabic, information, details, and rights.
Export the final version: At the quality and dimensions suited to your use.

If the result is poor, do not immediately switch to a new tool. First ask: is the problem the tool, or unclear goals, references, and dimensions?

Try it yourself

Lecture task: create an image you will actually use

Make at least one image for a real need in your work, studies, or life. It could be an educational design, infographic, video cover, carousel, product image, short illustrated story, or even a design for a phrase you like.

Steps to follow

Follow the workflow above for one image, from defining the goal to exporting.

What will you submit?

The final imageThe prompt usedThe reference, if you used oneThe first edit you made to improve the result

Summary

Start with one image you actually need

Start with a simple description. When you need more control or repeatable quality, add details, use a suitable reference, and edit one part at a time.

When you learn to make a good image and preserve its details, you build an important foundation for infographics, carousels, content designs, and later the videos we will work on.

I look forward to seeing what you create.
Stay Smarter With AI 💚