Just in
AIGPT-6.2 Astra and Sol Arrive as Astra API Prices Fall 20%CryptoBullish Earnings: Q3 Adjusted Revenue Up 10%, $300 Million BuybackTechNvidia Earnings: Q3 Revenue Up 103% to $115.8 BillionTechApple Outage: iCloud Sign In Issue Continues as App Store ClearsTechARTEX AI Agents Used in Key Theft and Deepfake Attacks, Amazon SaysTechNvidia National Security Board Committee Will Oversee Export Rules
AI · Guide

How to Use AI to Create Images, Step by Step With Gemini and Copilot

Free tools from Google and Microsoft can turn a sentence into a picture in minutes. Here is how to write the prompt, generate, edit and download your image.

USA Publishing Group6 min read
Share
Yellow 3D letters spelling AI on a dark surface with teal light, representing AI image creation tools
Photo: Jackson Sophat / Unsplash

Key takeaways

  • Google says Gemini users must be 13 or older to generate images and 18 or older to edit them, and you start from Open Sidebar, then Images.
  • A strong image prompt names the subject, the setting, the art style and the shape or mood, and small one change edits work better than many changes at once.
  • The U.S. Copyright Office says purely AI generated images are not protected by copyright, although your own edits and arrangement can be.
In this story 8 sections
  1. Before You Start Making AI Images
  2. Stage 1: Pick a Free AI Image Tool
  3. Stage 2: Write a Prompt the AI Can Follow
  4. Stage 3: Generate Your First Image
  5. Stage 4: Refine and Edit the Result
  6. Stage 5: Download and Use Your Image
  7. Who Owns What You Make
  8. When AI Images Do Not Come Out Right

You type a sentence, wait a few seconds, and a picture appears that never existed before. That is the whole promise of learning how to use AI to create images, and the free tools from Google and Microsoft now make it possible without any design skills or paid software. This guide is for anyone who wants a custom picture for a birthday card, a blog header, a class project or a social post. Your first usable image takes about 10 minutes, and most of that time goes into writing and refining the prompt.

Before You Start Making AI Images

A little setup saves frustration later.

  • An account. Gemini needs you signed in with a Google account, and Microsoft Copilot works best signed in with a Microsoft account.
  • The right age. Google's help page says you must be 13 or over (or the applicable age in your country) to generate images in Gemini, and 18 or over to edit them.
  • A clear idea of the use. A phone wallpaper needs a tall shape, a slide needs a wide one. Decide before you write the prompt.
  • Any photos you want to change. Microsoft asks that uploads be original images you own or have the rights to use, and Google warns against violating others' copyright or privacy.

Stage 1: Pick a Free AI Image Tool

Two of the easiest starting points are built into assistants you may already use.

  1. Google Gemini. Go to gemini.google.com and sign in. According to Google's Gemini help page, image generation runs on Google's Nano Banana 2 model, with a Lite version on the free Flash Lite tier and Nano Banana Pro available with a Google AI plan.
  2. Microsoft Copilot. Open the Copilot app or website and start a new conversation. Microsoft's support page says you simply type a prompt describing what you want, and Copilot takes a couple of minutes to produce the image.

Either one is fine for a first try. If you already pay for a Google AI plan, Gemini gives you sharper downloads, which matters if you plan to print.

Stage 2: Write a Prompt the AI Can Follow

The prompt is where most people go wrong. "A dog" gives you a random dog. A good prompt answers four questions.

  1. What is the subject? Name it plainly: "a golden retriever puppy."
  2. What is it doing, and where? "Sitting in a pile of autumn leaves in a backyard."
  3. What style should it look like? Microsoft notes that more detailed prompts help Copilot produce specific art styles such as realism, abstract, comic book or futuristic. Try "soft watercolor illustration" or "realistic photo, shallow depth of field."
  4. What shape and mood? "Wide format, warm late afternoon light, cozy feeling."

Put together, that becomes: "A golden retriever puppy sitting in a pile of autumn leaves in a backyard, soft watercolor illustration, wide format, warm late afternoon light, cozy feeling." For more prompt patterns, see our guide to writing better AI prompts.

Stage 3: Generate Your First Image

In Gemini

  1. On gemini.google.com, click Open Sidebar at the top left, then click Images.
  2. Choose one of the templates, or type your own prompt in the box.
  3. Submit and wait for the result.
  4. If you have a paid Google AI plan and want more detail, click More at the bottom right of the image and choose Redo with Pro. Google notes this option is blocked once you hit your daily quota of Nano Banana 2 images.

In Copilot

  1. Start a new chat in Copilot.
  2. Type your full prompt and send it.
  3. Give it a couple of minutes. Copilot shows the image in the conversation when it is ready.

Do not judge the tool by the first result. Generate two or three versions, then pick the one closest to what you pictured.

Stage 4: Refine and Edit the Result

This is where AI images go from "close" to "right." Both tools let you keep talking to the picture.

Editing in Gemini

  1. Click Open Sidebar, then Library.
  2. Click the image you want to change.
  3. Under the image, click Chat.
  4. Type the change in the text box at the bottom, such as "make the sky pink" or "remove the fence," and click Submit.

Remember that Google limits editing to users 18 and over.

Editing in Copilot

  1. Select the Open icon (the plus sign) and choose Add images or files.
  2. Pick the image from your device. It appears as an attachment.
  3. Type what you want changed, for example "remove the people in the background," and send it.
  4. Keep asking for changes in the same conversation until it looks right.

Change one thing at a time. When you ask for five edits at once, the AI often gets two right and quietly undoes something you liked. If you mostly want to fix real photos rather than invent new ones, our walkthrough on using AI to edit photos covers the phone tools built for that.

Stage 5: Download and Use Your Image

  1. Gemini: hover over the image and click Download full size. Google says downloads come in 2K resolution with a Google AI plan and 1K without one. You can also click More under the image and choose Export to Docs.
  2. Copilot: open the image in the chat and use the download option shown with it. Microsoft says generated and uploaded images stay available for a short time, no longer than 18 months, so save anything you care about.

Keep the prompt that worked in a notes file. When you need a matching image next month, you can reuse it instead of starting over.

Who Owns What You Make

This surprises many people. The U.S. Copyright Office has said that human authorship is required for copyright protection, and its 2025 report states that purely AI generated material is not covered and that prompts alone do not provide enough control. Your own edits, text and the way you arrange elements can still be protected. In practice, that means you can use these images for a newsletter or a flyer, but you may not be able to stop someone else from copying them. Businesses thinking about AI for branding should read our take on whether an AI logo generator can make a good logo before relying on one.

When AI Images Do Not Come Out Right

Most problems have a simple cause.

  • The tool refuses your prompt. Google says Gemini may remove images for prompts that break its terms or prohibited use policy. Prompts naming real people, famous characters or brands are common triggers. Describe an original subject instead.
  • Hands, faces or text look strange. Ask for a specific fix in a follow up message, like "fix the left hand so it has five fingers," or leave text out of the image and add it later in a photo editor.
  • You cannot find the Images option in Gemini. Make sure you are signed in, and check that your language and country are supported. School accounts can have extra limits set by the school.
  • You hit a limit. Daily quotas reset over time. Try again later, or switch to the other tool for the rest of your project.
  • The result looks fake when you wanted a real person. That is a different problem with its own tradeoffs, covered in our piece on AI headshots for LinkedIn.

Once you have a prompt you like, try changing just the style words. The same puppy as a pencil sketch, a 1970s poster and a clay model is the quickest way to learn what these tools can do. You can find more guides like this in our AI section.