Yes, AI can make a picture for you. Type a request like "make a picture for me of a red barn on a snowy hill at dusk" into an image generator such as ChatGPT, Gemini or Meta AI, and it draws a new image. It works from your description, so a request that leaves out the setting or the style gets a guess for both, and any spelled words need a check. Free plans cap how many pictures you make. Here is what AI can make, how to ask for it, how it edits your own photos, where it slips, and who owns the result.
What AI can make for you right now
AI can create images for you in several ways, and each one has its own kind of app. The main ones, as of September 2026:
| What you want | Where to do it | What to know |
|---|---|---|
| A brand-new image you describe | ChatGPT Images, Gemini, Meta AI, Toybox AI's Image Generator | Name the subject, setting, style and shape, or the AI picks them |
| An edit to your own photo | ChatGPT Images, Gemini, your phone's photo eraser | Say what must stay the same |
| A picture with you in it | Meta AI, Gemini, or reference photos in an image generator | The face can come out different from yours |
| A short video from a picture | Meta AI, Toybox's Video Creator | Gemini Omni, Gemini's video tool, and Video Creator need a paid plan |
| A diagram, chart or clickable visual | Claude, which builds them from HTML and SVG | Claude doesn't make photos or illustrations |
ChatGPT Images can also add text to a picture, add details inside it, or make its background transparent. Meta AI returns several images for one prompt, and adding the word "me" or "myself" to the prompt puts you in the picture. Claude's help center says it doesn't make photos or illustrations like an image generator. Its diagrams, charts and interactive visuals are built from HTML and SVG, a feature in beta on Claude's web and desktop apps. The details are in whether Claude can make images.
Logos and songs follow the same pattern with their own catches. Before you use an AI logo for a business, read can AI make a logo; for music, see can AI make a song for me. Video has its own answer in can ChatGPT make videos. To match a job to the right app, the job-by-job AI tool finder lists common tasks with the tools that do them.
How to get AI to make a picture for me
Describe the picture the way you'd brief an illustrator: what's in it, where it is, the light, the style and the shape. Put any words that must appear in quotes, and keep them short. Three requests that follow that pattern:
- "A watercolor of a red barn on a snowy hill at dusk, smoke rising from the chimney, two kids with sleds in the foreground, square, with 'Winter Open House' in large letters across the top."
- "A realistic photo of a golden retriever puppy asleep in a laundry basket of white towels, soft afternoon window light, shot from above, portrait 3:4, no text."
- "A flat cartoon map of a small park for a scavenger hunt: a pond, three benches, a playground and a big oak tree, bright colors, wide 16:9, no words."
In ChatGPT
Making a picture in ChatGPT takes five steps:
- Type the description in a chat, or select More, then Images, and enter your prompt.
- Choose a shape from the aspect ratio picker, or name it in your request ("square", "16:9").
- Wait while it draws. A detailed request can take a few minutes.
- To change one part, open the image, mark that area with Select, then type what should change.
- Choose Save to download it. Every picture you make is also kept under Images in ChatGPT.
In Gemini and Meta AI
Google suggests starting a Gemini request with a verb such as draw, generate or create, naming a style (watercolor, charcoal, cartoon), and describing the subject, what it's doing and the setting. To save the result, hover over it and choose Download full size. In the Meta AI app, open your request with "imagine" or the phrase "create an image"; Meta AI returns several images, and you pick one. For more on the ChatGPT side, see how to make images with ChatGPT.
Can AI edit my photo?
Yes. Upload the photo, say what should change, and say what must stay the same. That last part matters because AI editors can redraw more than the spot you mean: OpenAI notes that ChatGPT edits may reach beyond the area you select. The right app depends on the size of the change.
| The edit you want | What to type or do | Where |
|---|---|---|
| Remove a stranger or a trash can | Tap, brush or circle it | Your phone's built-in eraser: Clean Up on recent iPhones, Magic Eraser in Google Photos |
| Swap a dull sky | "Swap the overcast sky for a clear evening sky. Keep the people and the buildings the same." | ChatGPT or Gemini |
| Change a clothing color | "Make his shirt navy blue. Keep his face, hair and pose the same." | ChatGPT or Gemini |
| Tidy a messy room photo | "Remove the laundry pile and the cardboard boxes from the floor. Keep the furniture and the rug where they are." | ChatGPT or Gemini |
Phone erasers cost nothing extra. Clean Up works with iOS 18.1 or later on iPhones that run Apple Intelligence: the iPhone 15 Pro, 15 Pro Max, iPhone Air, and iPhone 16 or later models. Magic Eraser has been free for every Google Photos user since Google started rolling it out on May 15, 2024. The steps for each phone are in how to remove someone from a photo.
For the other edits, Gemini edits uploaded photos only for personal accounts aged 18 or older. Smart Editor in Toybox runs on OpenAI's GPT Image 2.5, takes one photo per run and needs a Pro plan. Each edit there redraws the whole image, so end every instruction with what stays, and switch between "Latest result" and "Original photo" to choose which version the next edit starts from. The full method, from choosing a tool to checking the result, is in how to edit a photo with AI.
Where AI pictures fall short
AI images fail in predictable places. Check these before you post, print or send one.
- Words in the picture. Signs, labels and headlines can come out misspelled. Check each word one letter at a time, and see why AI messes up text in images for fixes.
- Hands, faces and small objects. Count fingers, check eyes and teeth, and look at anything small in the background. Why AI is bad at hands covers the usual problems.
- Your likeness. A picture made from your photo is a new drawing, so it may not look exactly like you. A sharp, well-lit reference photo with your face clearly visible gives the AI more to work from.
- Other people. Google reminds Gemini users to respect other people's copyright and privacy, and Toybox's Terms of Service prohibit generating images of real people without their explicit consent.
- Free limits. OpenAI gives ChatGPT Free its own image limit, separate from chat, with slower generation than paid plans. Gemini saves images at 1K without a Google AI plan and at 2K with one.
- AI labels. OpenAI marks images made with ChatGPT with C2PA Content Credentials and SynthID watermarks. It says these signals help show where a picture came from but don't prove it's accurate or unedited.
Who owns a picture AI makes for you
In the US, a picture made entirely by AI from your prompt isn't protected by copyright. The US Copyright Office's January 29, 2025 report on copyrightability says copyright covers AI output only when a person determined enough of its expressive elements, and prompts alone don't count. Your own work that stays visible in the result, or creative changes and arrangements you make afterward, can still count.
The same report says that using AI as a helper, or placing AI material inside a larger work you made, doesn't block copyright for your own part. What you may do with the file depends on the terms of the app that made it. Toybox's Terms of Service say you own the rights to the outputs you generate (images, text, video, and code), subject to the Terms, and that you're responsible for how you use them. This is general information, not legal advice. The details for business use are in can AI generated images be used commercially.
Turn your picture into a short video in Toybox
If you want the picture to move, make the image first, fix it, then use it as the opening frame of a clip. Toybox's Image Generator makes the picture on the Free plan (currently 50 credits), and Video Creator, powered by Google Veo 3.1, animates it on a Pro plan (currently 350 credits per clip).
- In Image Generator, describe the picture and choose Wide (16:9) or Story (9:16), the same shape your video will be.
- Check the words, hands and faces, and download the PNG. On a Pro plan, fix small details in Smart Editor first.
- Open Video Creator, tap "Start from a photo" and add the picture as the start frame. Choose Wide or Vertical to match its shape.
- In "Describe your video", say what moves and what stays still, then pick a Length of 4, 6 or 8 seconds. Clips come with sound, so add the sounds you want, such as "soft wind and birdsong".
- Tap "Create video". It takes a few minutes.
- Watch the clip closely. Text and product labels can blur or change while they move, so keep the motion small.
A description that works with or without a start frame covers the scene, the motion and the camera:
A cup of coffee on a wooden cafe table by a window at sunrise. Steam rises slowly from the cup, soft morning light moves across the table, and a sheer curtain drifts in a light breeze. Slow push-in toward the cup, calm mood, no people, no text on screen.