There is a new kind of confusion around images. A client may say, “Can you retouch this?” But what they mean can be very different from traditional retouching. They may want the skin cleaned, the background changed, the light rebuilt, the fabric reshaped, or the image to look closer to an AI reference. Sometimes they want a real photograph to behave like something that was never photographed at all. That is where the conversation becomes more complicated. Not every image correction is the same type of work. Retouching is not the same as generating an image. Photography is not the same as AI production. Compositing is not the same as cleaning a file. And yet, in many conversations, all of these processes are starting to fall under one simple phrase: “Can you retouch this?” That phrase used to be more specific. Now, it often needs translation.
The image-making chain used to be easier to explain
For a long time, the production process was easier to describe. There was a shoot. A photographer created the image. The model, styling, makeup, lighting, location, and camera all shaped the photograph. Then the image moved into post-production, where retouching refined what was already there. Retouching could clean skin, adjust color, remove distractions, polish fabric, improve the background, and prepare the image for publication or commercial use. Of course, post-production was never simple. Compositing existed long before AI. Heavy transformations existed. Images that required serious reconstruction existed. But the language was still clearer. Photography meant that something was photographed. Retouching meant that the photograph was refined. Compositing meant that elements were combined or rebuilt. The borders were not always perfect, but each process had its own logic. Now those borders are softer.
AI changed the visual expectation
AI did not only add new tools. It changed what people expect images to be able to do. Clients and creative teams now see references where everything looks possible: perfect light, perfect skin, perfect fabric, perfect locations, perfect proportions, perfect atmosphere. No production limits. No weather. No location problems. No model fatigue. No physical constraints. Then those expectations are often brought back to real photographs. A client may show an AI-generated image and ask, “Can we make it look more like this?” Sometimes the answer is yes. Sometimes it is partly. And sometimes the real answer is that this is no longer retouching. It is a different kind of image-making process. That does not mean the request is wrong. It means the request needs to be named correctly. Because when a real photograph is expected to behave like an AI-generated image, the task changes. The time changes. The budget changes. The responsibility changes. The amount of reconstruction changes. And the conversation has to change too.
“Retouching” now means too many different things
One of the biggest problems is that the word “retouching” is now used for almost everything that happens after an image is created. A client may say “retouching” and mean skin cleanup, color correction, background cleaning, object removal, hair correction, clothing cleanup, body shaping, product polishing, image extension, background replacement, AI generation, compositing, or full visual reconstruction. These are not the same tasks. They may all belong somewhere inside post-production, but they are not the same kind of work. Cleaning skin is not the same as rebuilding light. Removing a small distraction is not the same as changing the environment. Adjusting fabric texture is not the same as making clothing look like it was photographed differently. The problem is not that clients always use the wrong words. The problem is that unclear words create unclear expectations. If everything is called retouching, then everything can also be expected to fit into the same timeline, the same budget, and the same level of responsibility. That is where misunderstanding begins.
Refining a photograph is not the same as rewriting it
For me, this is one of the most important distinctions. There is a difference between refining a photograph and rewriting it. Refining means working with what the image already contains. You improve the image while respecting its original logic: the light, the pose, the texture, the atmosphere, the person, the product, the intention. Rewriting means pushing the image into something the original file does not fully support. Both can be valid. Sometimes a commercial image needs heavy reconstruction. Sometimes a campaign needs compositing. Sometimes a product visual needs to be built from several images. Sometimes AI-assisted elements are part of the final direction. The issue is not whether rewriting an image is good or bad. The issue is whether everyone understands that it is a different task. When the request moves from refinement into reconstruction, the work is no longer just about making the photograph cleaner. It becomes about building a new visual reality. And that requires different planning.
A real photograph has physical logic
A photograph carries the logic of how it was made. The direction of the light. The quality of the lens. The depth of field. The pose. The fabric movement. The makeup texture. The way skin responds to light. The background perspective. The atmosphere of the shoot. Retouching can adjust many of these things. It can improve them, soften them, balance them, clean them, and sometimes reconstruct parts of them. But if every part of the image is pushed away from its original logic, the photograph can start to feel strange. The skin may no longer belong to the light. The background may no longer belong to the perspective. The fabric may no longer belong to the body. The color may no longer belong to the scene. This is why experienced retouching is not only about asking: Can this be changed? It is also about asking: Will the image still make sense after this change? That question becomes even more important now, when tools make more changes technically possible.
AI references can be useful, but they need context
AI references are not the problem by themselves. They can be useful for mood, color, styling ideas, atmosphere, and creative exploration. They can help a client communicate what they like when words are not enough. But an AI reference should not automatically become a production promise. A reference is not the same as a raw file. A generated image does not always follow the same physical rules as a photograph. It may have impossible light, inconsistent texture, unrealistic fabric, unusual anatomy, or a kind of perfection that was never created through a real shoot. When a client uses an AI image as a reference, the question should not only be: “Can we make it look like this?” The better question is: “What exactly do we want from this reference?” Is it the color? The mood? The softness? The styling? The background? The level of polish? The overall feeling? Once that is clear, the retouching direction becomes more realistic. Without that clarity, the retoucher may be asked to chase something that was never properly defined.
Why this matters for time and budget
This confusion is not only theoretical. It affects real work. If a client thinks a task is “just retouching,” they may expect it to be fast. If they see AI tools remove objects instantly, they may assume every correction should be simple. If they see generated images with perfect conditions, they may expect the same level of control from a file that was created under very different circumstances. But professional post-production is not only about making a change. It is about making the change work. If you remove an object, the background still has to make sense. If you change a shadow, the light still has to feel believable. If you clean fabric, the texture still has to look real. If you reshape something, the body, pose, and perspective still have to hold together. The visible action may be small. The consequences may not be. This is why naming the task correctly matters. It protects the client, the retoucher, the schedule, and the final image.
The retoucher is becoming a visual translator
I think the role of the retoucher is becoming clearer. Not because retouching matters less, but because the conversation around images has become more complex. A retoucher is no longer only asked to clean and polish. More often, they have to understand what kind of image is being requested in the first place. Is this a photograph that needs refinement? Is this a composite? Is this an AI-assisted image? Is this a full reconstruction? Is this a problem that should have been solved during the shoot? Is this a visual direction issue rather than a retouching issue? These questions are not separate from the work. They are part of the work. In that sense, the retoucher becomes a kind of visual translator. Someone who translates unclear requests into actual image-making decisions. Someone who understands the difference between a correction, a reconstruction, a reference, and a new visual direction. This is not only technical knowledge. It is visual judgment.
The industry needs better language around images
The future of retouching will not depend only on better tools. It will also depend on better language around images. We need clearer ways to describe what kind of work is being requested. We need to separate cleanup from reconstruction, refinement from generation, photography from AI production, retouching from compositing, and a visual reference from a production plan. Not because one process is better than another. But because each process has its own logic. AI can be useful. Retouching is still essential. Photography still has its own value. Compositing can solve complex visual problems. These things can work together. But they should not all be treated as the same thing. When the language is unclear, the work becomes harder to estimate, harder to explain, and easier to undervalue. When the language is clear, the process becomes more honest. And the image has a better chance of becoming what it actually needs to be.
