Wendy lost in the dreamy sequence of a jungle using Adobe Photoshop Generative AI

Adobe’s Brave New Generative AI World

Adobe is using Generative AI, a type of artificial intelligence that creates new data, such as images or text, by learning from existing data. It visualizes and generates content based on your descriptions and the existing content in a photograph or drawing, pushing boundaries for creatives and enthusiasts.

Photography and More

The photograph, Greek for drawing with light, is losing its purity with the onslaught of Generative AI apps that now speak of photos or pictures as “AI generated imagery” or synthetic media.”

Welcome to Adobe’s new world of Generative AI (Artificial Intelligence) where your favourite travel photo scanned from your film camera, digital camera or smartphone can become an image masterpiece in minutes.

Text prompts of a bunny playing a flaming saxophone in Adobe Firefly AI engine produced four samples

Firefly, is Adobe’s generative AI engine, often used in conjunction with the more general Adobe Generative AI name. It is in Adobe’s subscription Creative Suite and more affordable apps like Adobe Express on desktops, iPhones, and Androids. It also works on its own as Adobe Firefly the web app uploading and downloading photos, illustrations, and fonts on any browser. Unlike the previous above apps where its embedded, Firefly on the web has to download it’s generative wonders to other apps like Adobe Express for traditional editing and further use. But it has a rich interface.

Here is what the Firefly AI engine can do with various Adobe apps.  

  • Generative AI Engine: Firefly is a powerful generative AI engine available in Photoshop and Illustrator. It transforms the way you create by enabling you to generate images, add or remove objects, transform text, and more using simple text prompts in over 100 languages.
  • Generative Fill: Use a brush to remove objects or paint in new ones.
  • Text to Image: Generate images from detailed text descriptions.
  • Text Effects: Apply styles or textures to words and phrases.
  • Generative Recolor: Create colour variations of your vector artwork.
  • 3D to Image: Generate images from the interactive positioning of 3D elements.
  • Project Stardust: Move any object anywhere in your image.
  • Generative credits Allow access to these features in Adobe Photoshop, Illustrator, and other apps within Creative Cloud.

My Photoshop Wendy in Hydra Project

The original photo of Wendy in Hydra top left, had a Photoshop Generative makeover with backgrounds generated by text commands

My favourite all time travel photo of Wendy at the harbour of Hydra Island in Greece, originally shot on colour negative film, was due for an AI makeover. Not Wendy, of course. She remained the same as she was magically flown worldwide – even outside of it – for some virtual shoots. After scanning the negative, here is the work sequence in Adobe Photoshop version 25.5.1 armed with Generative AI.  

Generative Expand: Expanded the vertical photo beyond its borders by dragging the Crop tool beyond the edges of the original and typing “expand” in the switched-on Generative AI contextual working box. The expanded canvas automatically filled with content that seamlessly blends with the existing image. Note the natural extension of the mountains and water.

Adobe Photoshop Generative AI adds new elements like the lighthouse via text commands

General Fill: I selected two areas in the photo, one on the upper left and typed “small flock of birds flying.” I was given three options from which to choose. I could ask for more choices, but the first one was perfect. Another selection on the lower right side for “red lighthouse with shining light” yielded the perfect lighthouse which I reduced in size. It fit seamlessly.

Time for a new background. I removed the lighthouse and birds by roughly selecting both and typed “remove.” I then selected Wendy, the setting sun, and small rowboat, then reversing the selection for Photoshop to know I wanted to make changes on everything but Wendy, sun and boat.

Changing backgrounds from the original top, with text prompts like Jungle, sand dunes and stars

Content Fill: I simply typed “jungle” in the contextual box and pressed generate for a selection of jungle backgrounds. I repeated the Generate several more times until I got the perfect shot. Then I backed up to the third step for a new background and typed “sand dunes”. Note how Generative AI created a pond in the bottom of the jungle and sand dune scenes for the rowboat.

Typing backgrounds like “marsh and reflection” yielded a number of entertaining results. Further exploration typing “earth and moon” yielded a stunning surreal image of Wendy on a crusted lava hill, presumably a post-apocalyptic world against a giant nearby moon and a few new lonely planets afar.

Typing earth and moon generated a varying selection, but this one was a showstopper!

For my next adventure, I returned to the original wide view of Wendy but selected her and the pier she was standing on and generated more interesting scenes where the pier took odd shapes.

Continuing from the previous step with Wendy standing on an enlarged pier against the setting sun, I expanded the image to twice the size all around using the crop tools and typed even more imposing scenes like “huge waves” and finally “huge waves and thunderstorm.”

More than 10 Adobe Photoshop Generative AI stages typing brief text suggestions like thunder clouds and large waves

I was blown away with the realistic photo reality of Photoshop Generative AI with the final images rendered to more than 90MP, enough for a room-size wall mural. Am I now destined to put my camera aside to produce pretty images?

Worldwide photo contests have been blindsided, initially by winning entries having drastically edited entries. Some refuse such entries or cancel awards while others have created new categories for digitally altered photo-realistic images, easily done with the power of Generative AI in seconds.

Adobe has taken several steps to counter concern and criticism over its use of Generative AI like proposing the Federal Anti-Impersonation Act (FAIR) to protect artists from economic harm caused by AI impersonation of their style.

It also is working on a universal Content Initiative “Do Not Train” Content Credentials tag to promote accountability and transparency in generative AI.

The AI processing is done off site on an Adobe server. With my Intel i7 equivalent Asus 14 OLED laptop idling until the online rendering was downloaded, then starting some serious processing to show the final results in several seconds.

FYI, a smartphone-sized 12MP image takes about 15 seconds to render with a good internet connection and several seconds to process the returned image. Throughout, the app provides lots of help as you discover your way through.

Generative AI in Other Adobe Apps

Here is a summary of Adobe’S Creative Cloud Generative AI in other Adobe apps.   

Adobe Illustrator

Generative Recolor creates color variations with text guidance by user

Adobe Illustrator lets you create digital content in drawings and illustrations from scratch or using a plethora of premade files you can use in your projects. It uses raster graphics made up of paths, each with a mathematical formula called a vector that defines its shape and colour by moving long vector arms.  

Each final project can be revisited easily changing its characteristics like shape and colour. Unlike photos limited in blow-up size determined by the original camera or scanner that created it, raster graphics have no magnification limit. Generative AI comes into play when using Text to Vector Graphic producing graphics based on your text input within Illustrator.  

Generative Recolor creates color variations of your vector artwork based on text requests from mood to what colours you prefer to include or leave out quickly showing samples to pick from.  

Adobe Acrobat and Reader

Acrobat’s generative AI features, some in beta, use AI Assistant to quickly scan through lengthy, complex PDF content and get answers or generate a summary using Generative Summary helping improve comprehension, gain insights, and produce high-quality content quickly.

Project Stardust

Project Stardust Moves any object anywhere in your image.

Export Options

Firefly-generated content can be exported in various formats, including AI, SVG, and PSD depending in how you want to print of publish online. 

Generative AI for Fonts

Adobe Generative AI for eye candy fonts

Firefly allows you to generate stunning fonts from text descriptions. Describe the font style you envision, and Firefly will transform your words into unique and expressive typefaces.

Experimentation and Iteration

Firefly makes experimentation and iteration easy. Each Text Effect you generate produces multiple results, giving you different versions from which to choose. You can fine-tune how closely you want the effect to match the shape of your letters and play with fonts and background colors.

Availability and Cost of Adobe Firefly

Adobe Creative Cloud, Adobe Stock and Adobe Express have Generative AI embedded. Adobe Firefly on the web runs by itself. But the one thing they have in common is requiring you to use generative credits for running Adobe Generative AI. Here is a breakdown how AI works with generative credits for each Adobe app. Generative credits are needed to run in Adobe apps after their free beta ends. Some come free, usually 25 monthly for the free version of apps but offer more like 100 credits for a monthly premium fee. Generative credits used is based on how complicated the AI is. If you ask an AI project you started to generate more choices this will incur more Generative AI credits. Free credits also limit the size of your working image to a lower web quality 2,000 x 2,000 pixel size photo and include a Firefly waterboard marking on the photo.

Adobe Express Premium

Adobe has released a new app for Adobe Express Premium, its well designed cloud-based mobile design platform, bringing the same creative, editing, and Firefly-powered generative AI features enjoyed by desktop users to iOS and Android devices.

Impossible images generated by text commands with Adobe’s Firefly AI engine.

Here’s what Firefly does in Adobe Express. With simple text prompts in over 100 languages, you can generate images, add or remove objects, transform text, and more. Features include:

  • Generative Fill: Use a brush to remove objects or paint in new ones.
  • Text to Image: Generate images from detailed text descriptions.
  • Text Effects: Apply styles or textures to words and phrases.
  • Generative Recolor: Instantly reate color variations of your vector artwork.
  • 3D to Image: Generate images from the interactive positioning of 3D elements.
  • Project Stardust: Move any object anywhere in your image.

The free Express starts with 25 Generative Credits while the Premium plan has 250 monthly generative credits and access to Adobe fonts and no watermarks on images generated by Firefly for $10 a month.

Adobe Creative Cloud

The Adobe Creative Cloud subscription for the whole suite includes 1,000 monthly non-transferable Generative Credits.  If they run out, you can continue using the features but with slower results until the next monthly renewal.

Users with a subscription to a single app (such as Illustrator, InDesign, Photoshop, Premiere Pro, AfterEffects, Audition, Animate, or Dreamweaver) get 500 generative credits per month, while Adobe Stock paid subscriptions offer 500 generative credits monthly.

Adobe Firefly

Adobe Firefly on the web users get 25 free generative credits while Adobe Firefly Premium users get 100 generative credits for $10 monthly.

Find out about the Adobe AI compliance details here.