From Photoshop to Video and Audio, Adobe's New World of Artificial Intelligence
In the early days of visual production with artificial intelligence, the creative process was quite simple: We would write a prompt, wait a few seconds, and choose one of the images we came across. However, the need of the professional design world has never been just “create me a beautiful visual”. Designers need to enlarge the image, remove an object, expand the background, change a certain part of the product, adapt it to a different format, turn it into a video, and do all this without losing creative control.
The point Adobe Firefly has reached in 2026 is trying to solve exactly this problem. Firefly is no longer just a product that can be described as Adobe's AI image generator. Adobe positions the platform as Creative AI Studio, which combines image, video, audio and vector production, as well as professional editing tools, different artificial intelligence models and increasingly agentic AI features in a single creative working environment. Within Firefly, it becomes possible to create an image, edit the existing photo, increase its resolution, remove the object, expand the frame, produce video, prepare sound effects and even produce music.
This change is important. Because rather than trying to replace artificial intelligence with Photoshop or Illustrator, Adobe is making AI one of the new interfaces of the creative production process. Prompt can now be used not only to generate images from scratch, but also to replace existing creative and speed up professional workflows.
What is Adobe Firefly? Why Are We Talking About a Different Platform in 2026?
When Firefly first appeared, its most talked about feature was its ability to create images from text. Today, Adobe's definition of Firefly is much broader: a system that combines image, video, audio and vector tools in the same creative AI studio.
The real transformation here is the disappearance of the border between “generate” and “edit”. Once you create an image with AI, you don't have to go to another program and edit it. In the same working flow, you can create the image and replace a specific object, remove unwanted elements, enlarge the background, or move the image to a higher resolution. Firefly Image Editor, which Adobe expanded in March 2026, includes tools such as Generative Fill, Generative Remove, Generative Expand, Generative Upscale and Remove Background.
This approach targets one of the biggest shortcomings of generative AI for the professional user: control. Producing a great AI image from scratch is one thing, but being able to change only the desired region in a customer-approved image is another problem. The second problem is exactly the area in which Adobe has been specializing for years with Photoshop and Creative Cloud.
Editing the Photo Instead of Reproducing it: Firefly AI Photo Editor
The innovation seen on the second screen you sent is one of the clearest examples of this. Adobe now offers an “AI-supported photo editor” directly to the user in Firefly. The user uploads his/her existing image and can make changes to the photo with simple text commands.
The important thing about this system is that it not only produces new content with AI, but also brings the existing visual editing process closer to natural language.
For example, changing the background in a travel photo, removing an unnecessary object in a product photo, or recreating a specific part of the composition may require a lot of selection, masking, and editing with traditional tools. In the generative editing approach, the user can start the process by describing the result he wants.
In Adobe's current Firefly Image Editor, it is possible to add new elements to an area or change existing elements with Generative Fill, to remove unwanted objects with Generative Remove, to create outside the current borders of the image with Generative Expand, and to separate the subject from the background with Remove Background.
Here we see one of the most valuable aspects of generative AI in terms of professional design. The aim is not to tell the designer "you don't need to learn Photoshop anymore"; Leaving more time for the designer's actual decisions by delegating some of the dozens of small technical steps to AI.
Enlarging the Resolution of the Photo by 2x and 4x with Firefly
The first screen you sent shows Firefly's Upscale Image feature, which can be very useful especially for photographers, social media teams and designers.
The user uploads a photo and can enlarge the image 2x or 4x. However, we are not talking about a classic “image resize” process here. Adobe positions this system as Generative Upscale and includes partner models in addition to Firefly's own model.
Currently, Firefly Upscaler, Topaz Bloom and Topaz Gigapixel options can be used in Firefly. The user can choose 2x or 4x scale. By adjusting the "Creativity" level on the Topaz Bloom side, a result that is more faithful to the original image or produces more new details can be preferred. When using Topaz Gigapixel, there is a Face Recovery option to help preserve the details and expressions in blurry or low-resolution human faces.
This detail is especially important. Because in AI upscale systems, when the image is enlarged, the model sometimes reconstructs the missing details by guessing them. While this may seem quite successful for texture on a leaf, even a small guess on a human face can make the person look different. That's why Adobe's offer of different upscale models and controls such as Face Recovery makes more sense than just the "enlarge the photo" feature.
A low-resolution product photo, an old creative, a small image taken from social media, or a work created by AI at a lower resolution can thus be prepared for larger uses. There is a similar logic in Firefly's Fast Mode approach: it is possible to produce low-resolution variations for quick ideation and then upscale the desired result.
Creating Non-Existent Parts of the Photo with Generative Expand
Another extremely useful feature is Generative Expand. Imagine you have a horizontal photo but you need vertical creative for Instagram. With the traditional method, you may have to crop the image and lose important parts of the composition.
Firefly can do the opposite: instead of shrinking the photo, it enlarges the canvas and uses AI to create parts of the image that were never taken.
In Adobe's current tool, ready-made ratios such as 1:1, 4:3, 16:9 and 3:4 can be used in addition to the Freeform option. While you can describe how the new area should look with the prompt, you can also leave the area blank and allow Firefly to analyze the existing photo and create a continuation.
This feature is more important than expected, especially in advertising. Because a single main image needs to be used in different proportions, such as an Instagram post, Story, web banner, YouTube cover or horizontal ad format. Here, AI facilitates the adaptation of existing creative to different media formats rather than producing a new work of art.
Adobe No Longer Relying on a Single AI Model
One of Firefly's most interesting changes in 2026 actually happens behind the user interface. Adobe is not content with just offering its own Firefly models.
In the Firefly creative studio, in addition to Adobe models, models from Google, OpenAI, Runway, Black Forest Labs, Kling AI, Luma AI, ElevenLabs and other technology companies can be used in different workflows. While Adobe talked about access to more than 25 AI models in Firefly Image Editor in March, in August it defines Firefly as a studio that combines different industry-leading models in the same environment on image, video and audio.
This approach is a pretty smart change in Adobe's AI strategy.
Because for creative professionals, the question "which company's model won?" is not always the most important question. One model may be better at producing realistic people, another may be better at product visuals, and another may be better at video or upscaling. What is important for the user is to be able to use the model that suits his needs.
This is where Adobe begins to position itself not just as a model maker, but as a professional working environment for creative AI models.
In our opinion, this may be exactly one of Firefly's strongest advantages in the long term.
Firefly Now Turns into a Video Studio
The transformation in Firefly is not limited to photos. In addition to producing videos from text and images within the Adobe platform, it also offers web-based video editing features. The user can create clips via prompt or visual, work on existing video, and incorporate new AI productions directly into the video project.
At this point, Firefly sits wider than the classic text-to-video tool. Adobe's goal is not just to offer a "produce 10-second AI video" button; Creating a workspace that extends from production to editing.
For example, a creative team preparing a commercial can first create concept images with Firefly, animate these images with image-to-video, produce new B-roll footage, and then combine them in the video editing environment.
This approach is more realistic for the future of generative AI in professional production. Because a real commercial does not consist of just a single AI clip. In addition to the image, editing, sound, music, speech, formatting and revision are required.
Adobe is trying to gather exactly these parts under the same roof.
Sound Effects, AI Dubbing and AI Music Now in Firefly
One of Firefly's most recent expansions in August 2026 was on the audio side. The platform now combines image and video, as well as music, speech and sound effect production, within the same creative studio. In the current structure of Firefly, Adobe uses partner technologies such as ElevenLabs in addition to its own tools.
In Firefly's video editing tools, audio can be produced in more than 20 languages with text-to-speech. Being able to adjust features such as emotion, tempo and emphasis can be especially useful in advertising, training and social media videos.
One of the more interesting features is Generate Sound Effects. You can describe the sound you want with text, upload a reference sound, or imitate the sound effect yourself using a microphone. Firefly can use this as a reference for the effect to be produced. For example, it becomes possible to describe in words the sound of a fantastic creature's wings or a particular automobile passing effect.
Generate Soundtrack aims to analyze the story of the video and create original music suitable for it. This reduces the need to switch to a completely different AI platform for music after rendering the image.
This is where the true meaning of Firefly's "all-in-one creative AI studio" claim emerges: Create the image, convert it to video, edit it, voice it, add effects and produce your music.
AI Avatars Are Also Included in Adobe Firefly
Another rapidly growing area in corporate video production is AI avatars. Adobe Firefly also enters this field with its Text to Avatar feature.
A realistic digital presenter that speaks via text can be created. This type of technology is especially important for internal training videos, product descriptions, social media content, onboarding materials and multilingual corporate communications.
Thus, some types of content that require cameras, studios, servers and reshoots for each minor update can be produced faster.
It is not right to think that this technology will completely replace traditional production. Brand film, emotional advertising or high production value storytelling is another area. However, in corporate structures where hundreds of training videos need to be updated regularly, the economic value of AI avatars becomes much more understandable.
Firefly Custom Models: Can AI Learn the Visual Language of a Brand?
One of the biggest problems of Generative AI for brands is consistency. The first image may be perfect, the character changes in the second image, the color world is distorted in the third production, and the tenth creative no longer looks like it belongs to the same campaign as the first work.
Adobe is developing the Firefly Custom Models approach for this. The user or brand can create a customized model according to a specific creative style by using their own images. The system, which Adobe expanded as a public beta in March 2026, focuses especially on ideation processes in character, illustration and photography styles.
This technology is perhaps one of the most critical areas of generative AI for agencies and big brands.
Because in the future, the question will not only be “Can AI produce beautiful visuals?” It won't happen.
“Can AI produce visuals like our brand?” The question will be much more important.
When a fashion brand's lighting approach, a cosmetics brand's product shooting aesthetic, or a game studio's character world can be consistently maintained by the model, AI can transform from a one-off production tool into a true brand production system.
Firefly AI Assistant: Moving from Giving Prompts to Getting AI to Do the Work
Perhaps Adobe's biggest innovation in 2026 is not a single visual feature: Firefly AI Assistant.
AI Assistant aims to have the user describe the desired result in natural language and have Adobe's creative agent system perform the necessary steps. Adobe describes this as a model where the creative professional remains in the director's chair while AI organizes complex and repetitive steps.
The difference is quite big.
In today's AI usage generally:
“Create this image.”
we say.
In Agentic systems:
“Adapt this campaign visual to Instagram Story, feed and web banner formats, expand the backgrounds and prepare different formats.”
The aim is to give a result-oriented instruction such as.
AI is starting to organize which tool will be used in which order.
As of June 2026, Adobe offers AI Assistant as public beta in Premiere, Photoshop, Illustrator, Frame.io and InDesign, while running private beta on After Effects. Firefly's own AI Assistant features also continue to expand in beta.
This change is one of the clearest examples of the transition from generative AI to the agentic creativity era.
Photoshop, Premiere, Illustrator and InDesign are also included in this transformation
Adobe's advantage is not that Firefly is strong on its own. It is backed by the Creative Cloud ecosystem, which has been used by the professional creative industry for decades.
Photoshop is among the important professional tools for image editing, Illustrator for vector design, Premiere for video editing, After Effects for motion design, and InDesign for publishing. Adobe is now expanding its AI Assistant approach into these applications.
Therefore, it is incomplete to consider Firefly as a separate AI site.
In the system that Adobe is trying to establish, Firefly AI can become the production and guidance center, and Creative Cloud applications can become the professional control and final production layer.
An idea can start in Firefly, be detailed in Photoshop, create vector elements in Illustrator, turn into video in Premiere, and distribute to social media formats through Express.
The real power is not in a single AI feature, but in the interconnection of the entire chain.
Adobe Brings AI Tools to ChatGPT, Claude and Gemini
What's even more interesting about Adobe's strategy is that it doesn't expect users to always come to the Adobe application.
The company is also bringing its Creative Cloud tools to third-party AI platforms. Adobe announced in June 2026 that its tools are available on platforms such as ChatGPT, Claude and Microsoft 365 Copilot, and that it is also working on connections for Google Gemini and Slack.
On Adobe's current Firefly page, it is stated that creative Adobe tools can be used within Claude; The user can edit multiple photos or adapt videos to different platform formats without leaving the chat environment.
This gives us an interesting clue about the creative software of the future.
Maybe in a few years, instead of saying "Open Photoshop":
“Clean the background of this photo, preserve the product, create 4:5 and 9:16 versions, and increase the resolution.”
We will say.
Photoshop, Firefly or other Adobe services will run in the background, but the user will not have to think about which button of which tool to press.
Why Is Adobe Firefly Particularly Important to Designers
There are hundreds of platforms in the world of Generative AI. A new image or video model comes out every month. That's why Firefly's success is based solely on the question "which model produces the most realistic photo?" It may not be correct to connect it to the question.
Adobe's strength is workflow.
A designer's need is not just to produce images. It is necessary to edit the image, enlarge it, adapt it to a different format, change the background, work with layers, add typography, prepare it for printing, apply the customer's revision and deliver the final file.
The first generation of Generative AI tools were only very powerful at the beginning part of the creative process.
Adobe, on the other hand, is trying to gradually spread AI throughout production.
Moreover, the fact that it brings together different models in Firefly makes the question "Which is better, the Adobe model, the Google model or the OpenAI model?" It may make the discussion less important to the user. The right question for the designer becomes:
Which model works better for this job and where can I edit the result professionally?
Firefly is trying to answer these two questions within the same platform.
Will Adobe Firefly Replace Photoshop?
In our opinion, no; At least that's not what Adobe's strategy looks like today.
Firefly is becoming a new way to access some of the operations that Photoshop can do, rather than being an alternative to Photoshop. Photoshop provides much deeper control over precision selections, layers, masks, color management, compositing and professional retouching. Firefly, on the other hand, uses natural language and generative AI to start many processes faster.
Therefore, in the future, the relationship between the professional designer and AI will probably be “AI or Photoshop?” It won't be like this.
More correct question:
“How fast can Photoshop + Firefly work together?”
It will happen.
With Firefly, a creative professional can create 20 concepts in minutes and take two selected ideas in Photoshop to a professional level. You can prepare different formats with Generative Expand, increase the resolution with Generative Upscale, and then create an animated version of the same campaign with Premiere and Firefly video tools.
This is a scenario where AI expands the designer's production capacity rather than a scenario where AI eliminates the designer.
Adobe Firefly's Biggest Innovation in 2026 Isn't Actually a Single Feature
It's nice to enlarge the image 4x.
It's nice to edit photos with Prompt.
It's nice to remove an object with a single command.
Producing a video from a photo, performing AI voice-over or automatically creating music for the video is also impressive.
But in our opinion, the real important innovation of Adobe Firefly is none of these.
The real innovation is that all these become parts of the same creative ecosystem.
In the early days of Generative AI, creative professionals were navigating through dozens of tabs. He was creating images on one platform, upscaling them on another, producing videos elsewhere, preparing voice-overs on a different service, and then moving all the files to Adobe applications.
Adobe is now trying to reverse this fragmented process.
Firefly; It becomes the creative center where images, video, audio, vector, AI models and agentic tools meet. Photoshop, Illustrator, Premiere, After Effects, InDesign, Express and other Creative Cloud applications constitute its professional production infrastructure.
This is exactly the main argument that Adobe used when adding music, speech and sound effects tools to Firefly in August 2026: when a creative idea arises, switching between different applications, tabs and models slows down the creative flow. Firefly aims to collect these pieces in the same studio.
For this reason, we find the point Adobe has reached on the artificial intelligence side to be quite successful. Rather than just joining the AI race with a vision model, the company began using the real advantage it's had for years: knowing how creative professionals work.
And perhaps Adobe's biggest trump card in the AI race is right here.
Midjourney can produce a more impressive visual. Another model can create more realistic video. Another platform may produce better sound. These rankings will constantly change as models evolve.
But for the professional creator, the work does not end when the visual is produced.
Most of the time, on the contrary, the real work begins at that point.
What makes Adobe Firefly's transformation in 2026 important is that it begins to transform artificial intelligence from being just a "producing" technology into a working system that can be used from the beginning to the end of creative production.
If Adobe can make this integration as seamless as it promises, we may not remember Firefly as "Adobe's AI image generator" in the future.
Instead, Firefly could be the centerpiece of the transition from the Creative Cloud era to the Creative AI era.
Nur Oğuz