When people talk about AI-generated work, cost almost never comes up. Everyone sees the output; nobody sees what happens behind the curtain.

This isn’t actually a new problem.

It was here long before AI. “What’s the big deal, you’re just sitting at a computer,” people would say. The invisibility of digital labor is an old wound. When there’s no physical product at the end of it, the effort never quite feels “real” to some people. AI deepened that perception, because it added one more thing to the mix: speed.

It was made fast.

So it must be easy. It was easy. So it must be cheap.

That chain of logic is everywhere — and it’s completely wrong.

Think about it: these platforms aren’t free. Behind them sit enormous servers, serious energy consumption and real infrastructure. They pay a price to be able to offer that service, and they pass it on to us.

Nothing in life is free, and AI is no exception.

And there’s this too: you don’t get the result you want on the first prompt.

You change the light, you change the material, you change the angle. Every generation burns credits. The dollar you spend on one image at the draft stage is only the beginning.

Then the client revisions arrive. You type the same prompt and you don’t get the same result — consistency is its own labor, its own cost. By the end of the project the numbers pile up. With animation it grows exponentially.

Here’s the paradox: AI made the work faster but it didn’t make it cheaper. In some cases it made it more expensive.

Now add the professional’s expertise on top of that — the “eye” it takes to use the tool properly, and the years spent developing that eye. Those have a value too.

The phrase “made with AI” gets thrown around very easily now.

But nobody asks what sits behind that phrase, what it cost, who managed it and how. It’s exactly the right moment to start asking.

This perception is being shaped right now. If it settles on the wrong foundations, it will be much harder to correct later. Saying it out loud is in the interest of both the person doing the work and the person commissioning it.

“You’re using AI anyway, so why is this so expensive?”

Every digital creator knows exactly how it feels to hear that sentence. The first thing that comes to mind is: “If it’s that easy, you do it.” But instead you smile and start explaining the years behind the work.

This perception isn’t new. It was here long before AI. “What’s the big deal, you sit at a computer, click click click, it’s done in two minutes.” The invisibility of digital labor is a much older problem. AI only deepened it by another layer.

And there’s this: in Turkey, the bargaining culture is often built on devaluing the other person’s work. Make the product look worthless and you can push the price down. This isn’t specific to our industry — it’s everywhere. But it cuts sharper in creative and digital work, because from the outside there’s no physical product, no physical exhaustion, no crisis being managed.

And yet one fact doesn’t change: for that output to exist, thousands of projects had to come before it. Wrong decisions, attempt after attempt, nights spent staring at a screen, software costs, hardware investments — and most importantly, the years it took for that eye to reach this level. Whatever the tool is, that accumulation stays the same. AI speeds it up; it doesn’t replace it.

What is the client actually buying? A file? A few images? No. They’re buying the vision, the experience and the judgment behind that output. And that judgment didn’t come cheap, because it never was cheap.

In a period where economic conditions have been anything but easy, I know that our clients and we are both inside a difficult equation. Sometimes you have to meet in the middle. Sometimes flexibility is needed. But holding the line between flexibility and devaluation matters — for your own sake and for the future of the industry.

Every client you get to understand your value moves the industry to a slightly healthier place. A small difference, but an important one.

AI-generated architectural images get more impressive by the day. Sometimes it’s almost impossible to tell them apart from a real photograph at first glance. But here’s what I’ve noticed: no matter how beautiful an image is, if it’s technically wrong it’s completely useless.

Staircases that lead nowhere. Balconies with no railings. Ergonomically impossible spaces. Objects at the wrong scale. I see these every day. AI can produce an aesthetically satisfying picture, but it has no way of knowing whether the thing it produced could actually be built in the real world. Only someone who genuinely knows the field can catch that.

There’s an important distinction here: the intended use changes everything.

In images used at the concept or decision stage, these kinds of errors are tolerable. What matters there is conveying an idea, not technical accuracy. But when it comes to the sales imagery of a real project, the situation is very different. The client looks at that image and says “yes, this is what I want,” and makes an investment decision based on it.

That’s the point where a beautiful but technically wrong image can turn into a serious financial or contractual dispute between the developer and the client.

In my own workflow, AI plays a role in both the process and the output. First I use it to test framing and lighting alternatives — comparing morning, midday and evening light to see which one best captures the feeling I’m after. When a project asks for a night version, I use AI on top of the daytime render to explore ideas. In a way I use AI as a drafting tool — a way to test an idea before committing to it. But in the end I build every project’s 3D model myself, render it myself and check the technical accuracy myself. AI adds speed and aesthetic refinement; I build the foundation.

To everyone using AI as a tool: without a technically trained eye to verify what you produce, that tool will sometimes leave you with nothing more than a beautiful mistake.

When I was a small child, my grandfather recorded everything on video. Analog cameras, film sent off to be developed, hours spent outdoors… I owe him both my love of the visual world and my interest in architecture. I didn’t know it at the time, but in those years I was learning to read light and to feel a frame.

Then cameras went digital. Then 3D arrived. Now there’s AI.

The tools changed. But the grammar of seeing never did.

When I make an image, whether it’s a 3D render or an AI output, the real question is never technical perfection. It’s catching that shared feeling. The moment when the person looking can imagine themselves inside it and feels something they can’t quite put into words. Before I show anything to a client, I revise it for myself over and over.

Until I feel it.

Where does that feeling come from? From knowing what light falling on a surface at a certain angle is telling you. From sensing how depth of field loads a frame with emotion. From being able to read where a composition takes the eye. These are things photography, cinema and fine art have developed over centuries. AI didn’t wipe out that accumulation; it combined it with speed and flexibility.

Let me say this clearly too: this is not an “anyone can do it” story.

Prompt writing can be learned. But understanding why a particular prompt works, why light coming from that angle creates that emotion, why visual consistency across a series matters — that takes real knowledge and experience. And there’s something else. AI is fed by everything ever shared on the internet. A century of photographic theory, cinematic aesthetics, architectural vision, it’s all in there. But to reach that knowledge in a meaningful way, you have to ask the right question. And this is where something very human comes in.

Being able to direct AI well develops a skill that works in every area of life: asking the right question.

Knowing what you want. Defining your need clearly. Expressing yourself in precise language. Being able to ask for exactly the thing you need — from a person as much as from a machine.

AI is pushing us in that direction. It sharpens our ability to question, to think clearly and to hold a vision. That’s why the people who use AI well aren’t just tech enthusiasts; they’re people who know themselves, know what they want and know how to ask for it.

And finally, to everyone in this industry still asking “will AI take my job?”: the answer is that AI isn’t what will take it. It will be the competitor who learned to use AI while you were resisting. Resisting is a choice, but it has a price. Integrating this technology into the skills you already have is also a choice, and it makes you both faster and stronger.

Anyone who truly loves their work follows their industry, stays current and keeps growing. Not rigidity, but adaptation. Always.