Edit Template

Sometimes images don’t fit the required aspect ratios. Usually, social media posts, thumbnails, or web banners require you to adjust the photos. Many big enough images can often be cropped. But sometimes this is not the best option, and the result might not be usable. That’s why it’s better to use tools that can enlarge the image. In this tutorial, I will show you how to use AI to expand an image in Photoshop – Photoshop AI expand for web and desktop.

This article will explain what AI image expansion is and why it is useful. I will also provide two different tutorials. One for web-based Photoshop with the crop tool, and the second one for the desktop version by using the marquee tool.

Photoshop ai expand

What Is AI Image Expansion in Photoshop?

AI image expansion uses AI to generate new pixels beyond the original image’s boundaries. The tool is often called Generative Fill or Generative Expand, powered by Adobe Firefly.

When applying Photoshop AI expand, Photoshop analyzes existing image content, including lightning, textures, and perspective. Then it uses AI to create realistic extensions that can blend naturally with the original image. Before the existence of this Photoshop AI expand tool, people often used content-aware fill. Unfortunately, the results are not the best when dealing with more complicated scenes or compositions. Thanks to the Photoshop AI expansion, you can now let the AI invent believable new scenery, not just copy nearby pixels.

Why Use AI to Expand Images?

When compared to manual cloning or retouching, Photoshop AI Expand can save you a lot of time. It is an excellent tool for any web-based project that requires the image to fit specific measurements. For example, web headers, hero images, social media posts, print layouts, travel, and portrait photography.

Photoshop AI expand can preserve image quality and composition. It can minimize the need to re-shoot some photos. Thanks, the images are more flexible. You can adjust them to different platforms and layouts quickly with just a few clicks. The tools are suitable even for beginners.

When AI Image Expansion Is Useful

  • Expanding Photos for Web & Social Media
  • Fixing Tight Crops in Photography
  • Creating Banners, Thumbnails & Hero Images
  • AI Expansion for E-commerce Product Photos

What You Need to Use Generative Expand in Photoshop

To use the Photoshop AI expand tool, you need to work with the latest versions of Photoshop that support Adobe’s AI system.

Supported Photoshop Versions

To work with the latest versions, you will need to have an active Adobe Creative Cloud subscription and up-to-date Photoshop version (or at least the latest one that supports the AI generative activities)

Photoshop AI expand works best on the desktop app, not older or offline versions. Some of its features may not be available when you don’t update it.

Adobe Firefly & Generative Credits Explained

Adobe AI tools are part of the Adobe Firefly system. To use it, you need credits. A certain number of credits comes with your Adobe subscription (even the cheapest one, like the Photography plan), and the pack is around 250 credits. You can check your credit balance anytime by clicking on your account menu. Credits are shared across Adobe AI tools. Basic usage is enough for everyday editing workflows. If you use your pack of credits, the AI features should continue to work, but at a slower speed. Regenerating many variations or expanding large areas uses more credits. Heavy experimentation can use credits faster. One generation uses 1 credit and produces three versions to choose from.

Premium generative features are available only for specific plans. If you’re on the Photoshop Single App, Photography, or Creative Cloud Standard plan, some advanced features will be locked behind a paywall. You can check all the Adobe plans that include both standard and premium generative features on the Adobe site.

Some tools, such as Generative Remove in Lightroom or the Remove tool in Photoshop, don’t use credits.

Free members of Adobe Express, Adobe Firefly, and Creative Cloud receive a limited allocation of generative credits to explore and experiment with generative AI tools. The number of credits included with free plans may change.

How to Expand an Image Using AI in Photoshop (web-based)

Step 1: Open Your Image in web-based Photoshop on photoshop.adobe.com.

You can drag and drop the image or import it from your computer, phone, Lightroom, or Google Drive.

Step 2: Select Generative-> Generative Fill-> Generative Expand from the left menu.

Step 3: Extend the Canvas with the Crop Tool

Now you should see an image with crop tool marks around it. You can drag and adjust them to the size you want in any direction.

Step 4: Use Generative Fill to Expand the Image

If you are satisfied with the size, click the Expand button in the menu that should appear below your picture. If you want something specific, you can write it into the text field and then click on Expand. You can write the prompt in more than 100 languages. 

Step 5: Review and Choose AI Variations

Adobe Photoshop will generate three different versions. You can create more by clicking on the plus button next to those variations. If you click the three dots in the left corner of the menu, another menu will appear – there, you can choose Generate similar or Delete the variation.

You can also change the version of Adobe Firefly on the left side of this menu. By clicking on the arrows, you can move to the following variations or go back to previous ones.

Step 6: Export your image

Just click on the blue button in the top right corner. You can export the image as JPG, PSD, PNG, TIFF, or WebP. You can also download it on the phone by scanning a QR code.

How to Expand an Image Using AI in Photoshop (desktop)

Step 1: Open Your Image in Photoshop 

In the top menu, click on the File-> Open. Find the image you want to use for the Photoshop AI expand.

Photoshop AI Expand

Step2: Change the size of your canvas: Image-> Canvas Size. A new window will appear where you can change the size of your canvas. To do generative fill, increase the width, height, or both. Leave the anchor in the middle to evenly distribute the canvas. But feel free to experiment and change the anchor point.

Step 2: Choose the rectangular marquee tool from the left menu.

Choose the tool, then make a marquee around your image from the inside. Leave some space from it next to its corners. Then go to the top menu and choose Select-> Inverse. Now you should have selected the remaining expanded canvas.

Step 3: Go to the top menu and choose Edit-> Generative fill. A new small window will pop up. You can write your prompt into it or leave it empty. Then click Generate.

Photosop Generative Fill

Step 4: On the right, a new window named Properties should appear. If you don’t see it, go to Window-> Properties. There should be your first three generated variations. If you are not satisfied with the results, you can create more.

Step 5: Export the image by clicking on File-> Export. You can make a quick export as a PNG file, or use Export As or Export for web to adjust the image size or quality, and choose a different file format, such as JPG or GIF.  

AI Expand Photoshop

Best Prompts for AI Image Expansion

If you are not sure whether to leave the prompt field empty, here are some tips.

When to Leave the Prompt Empty

Photoshop analyzes existing pixels, lighting, perspective, and color tones. By leaving the prompt empty, it usually produces seamless backgrounds and a natural continuation of walls, skies, floors, and landscapes.

Best use cases

  • Extending skies, streets, beaches, forests
  • Fixing tight crops in travel or street photography
  • Expanding portraits without changing the mood
  • Expanding portrait framing for: Instagram, YouTube thumbnails, website headers
  • Creating room for text in banners or blog images

If you want the expanded area to look like it’s already been there, don’t tell the AI what to invent. Empty prompts can keep results more “photographic” than “illustrative”. Empty prompts will also rarely introduce new objects you didn’t want.

Tips for prompts

Use simple nouns + lighting, avoid adjectives like fantasy, dramatic, surreal, unless that’s your goal. Try to use fewer than 10 words – that often works best. Think of Generative Fill prompts as gentle suggestions, not instructions.

Empty prompt = safest & most realistic

Simple prompts = controlled realism

Detailed prompts = creative direction (but higher risk)

Generative Expand vs Content-Aware Fill

If the content doesn’t exist in the photo, AI Expand is the better choice. 

AI Expand can

  • Continue architecture
  • Guess scenery behind the frame
  • Match lighting and depth more naturally

AI generative fill understands context, not just pixels, so that it can invent believable content beyond the original frame. It is best when you need to add new space, not just remove something, and usually the background isn’t entirely present in the original image.

Photoshop AI expands alternatives

Photoshop’s AI Expand is powerful, but it’s subscription-based and sometimes too slow or too heavy for simple tasks. It doesn’t always fit every project, and some alternatives can produce faster results, be free, or offer different styles. One of such alternatives is the Hypic photo editor. It is a mobile app that includes tools for AI-generated images, including the AI Auto Fill feature. Although the free version offers less control over auto-fill, its results are impressive, engaging, and unpredictable. Besides the AI Auto Fill you can use the AI expand inside the Hypic Crop tool. It lets you insert your own prompt for the area you expand.

Final Thoughts on AI Image Expansion in Photoshop

Photoshop’s AI Image Expansion is best for reframing images, extending backgrounds, and fixing tight crops. It saves time compared to manual methods, and the results are often “good enough” straight out of the box. But on the other hand, it can misjudge details, add unwanted elements, and be a struggle when precision work is needed. Professional results still require human judgment, manual cleanup, and multiple generations. I have used this tool a lot to generate hands, but in the end, I was often forced to make manual adjustments to hide the AI’s unwanted or incorrect interpretations.

Sources:

Edit Template

Sources: