How to Generate AI Images: Prompts, Tools and Rights
Write image prompts that work, pick the right generator, edit results and understand commercial-use rules.
On this page
AI image generators turn written descriptions into pictures. They're useful for blog headers, social posts, mood boards and product concepts. This guide covers how to describe what you want, which tool to use, and what to check before you publish.
Step 1: Choose a tool for your purpose
- Quick images inside an assistant: ChatGPT and Gemini can create images in conversation, which makes iterating easy.
- Artistic, high-impact visuals: Midjourney is known for striking aesthetics.
- Photo editing and commercial design: Adobe Firefly, built into Photoshop and Adobe Express.
- Templates plus AI: Canva, if you need the image inside a finished design.
- Lots of creative control or offline use: Leonardo.Ai or Stable Diffusion.
Browse our AI image generators for details on each.
Step 2: Describe the image in layers
Good image prompts describe several elements:
- Subject: who or what, doing what.
- Setting: where, and what's around them.
- Style: photo, watercolour, 3D render, flat vector illustration.
- Composition: close-up, wide shot, top-down, centred, rule of thirds.
- Lighting and colour: soft morning light, high contrast, pastel palette.
- Mood: calm, energetic, professional.
- Format: aspect ratio, such as 16:9 for a header or 4:5 for Instagram.
Example: "Flat vector illustration of a young man studying at a desk with a laptop and textbooks, evening, warm lamp light, calm mood, simple shapes, navy and orange palette, 16:9."
Try our image prompt generator to build prompts like this.
Step 3: Generate several options, then refine
Generate a few variations and pick the closest. Then change one thing at a time: "Same image, but make the background plain white" or "Show the person from the side." Changing many things at once makes it hard to see what worked.
Step 4: Fix common problems
- Garbled text in images: many models struggle with lettering. Leave text out and add it afterwards in a design tool.
- Odd hands or faces: regenerate, use an edit/inpaint tool on that area, or choose a composition where they're less prominent.
- Too generic: add specific details – materials, location, time of day, cultural context.
- Inconsistent characters: use reference image features where the tool supports them.
Step 5: Edit and finish
Use editing features (inpainting, expand, background removal) to adjust the image, then place it into your design. Crop to the right size for each platform and compress the file for fast loading on websites.
Step 6: Check rights and responsibilities
- Read the tool's terms. Commercial-use rights often depend on your plan.
- Avoid real people's likenesses and copyrighted characters, logos or artworks.
- Label AI images where appropriate. Some platforms and laws require disclosure, especially for realistic images and political or news content.
- Don't create misleading images – for example, fake product results or fake events.
Next steps
Compare two popular tools in Midjourney vs DALL·E, and find ready-made prompts in our image generation prompts.
Frequently asked questions
Can I sell AI-generated images?
It depends on the tool's terms and your plan, and on the laws in your country. Copyright protection for purely AI-generated images is limited in some countries.
Why does text look wrong in AI images?
Many image models struggle to render letters accurately. Add text afterwards in a design tool.
AI Resource Hub Editorial Team
Our editors research AI tools using official documentation and reliable reporting, and update guides when products change. Read our editorial policy to see how we work. Editorial policy.
Tools in this guide
Midjourney
Midjourney, Inc.
Highly stylised AI image generation from text prompts.
- Text-to-image generation
- Image references for style and character
- Variations and upscaling
Adobe Firefly
Adobe
Adobe's generative AI for images, effects and video, built into Creative Cloud.
- Text-to-image
- Generative Fill and Expand in Photoshop
- Text effects
ChatGPT
OpenAI
General-purpose AI assistant for writing, analysis, images and conversation.
- Chat with text and voice
- File and image uploads for analysis
- Web search with linked sources
Leonardo.Ai
Leonardo.Ai (Canva)
Web-based image and video generation with fine creative controls.
- Multiple image models and style presets
- Canvas editor with inpainting
- Custom model training
Canva
Canva
Online design platform with AI features for graphics, slides and video.
- AI image generation
- Background remover
- Resize designs for different formats
Related articles
Midjourney vs DALL·E (ChatGPT Images): Which Image Generator Wins?
Compare Midjourney with OpenAI's image generation in ChatGPT on quality, ease of use, editing, text in images and pricing.
2 min read
How to Use AI for Social Media Content
Plan, write and design social media content with AI while keeping your voice, accuracy and audience trust.
2 min read
How to Use AI for Your Small Business
Practical, low-cost ways to use AI for marketing, customer service, admin and planning — with guardrails.
2 min read