Will AI replace designers?
Execution got cheap. Taste did not. The problem is that many designers were selling execution while believing they were selling taste.
A small confession first.
I used to spend a long time on thumbnails for my videos. Long enough that I seriously considered hiring a graphic designer.
Then I stopped considering it entirely, because one instruction started giving me what I wanted.
And when tools like Nano Banana and GPT Image got good at Arabic specifically, the argument ended for me. I no longer see a reason to hire anyone for the simple work.
I say that as someone from this field: I worked as a creative director and a cinematographer before I trained anyone. So I am talking about colleagues, not strangers.
What changed is not design. It is the price of execution. And that price did not fall. It collapsed.
What got cheap
First-round concepts. Twenty versions of a banner at different sizes. Background removal. Colour variations. A stock-style image instead of a licence. An interface mockup.
All of that used to be billed by the hour, and most of it was not design. It was execution.
What did not get cheap
Visual judgement. Looking at thirty outputs and knowing which one works and why. The system generates; it does not discriminate. Its capacity to say "that is a cliché" is zero.
Understanding the problem. Most client briefs describe a solution rather than a problem. "I want a blue logo" is not a problem. Getting to the problem is the job.
Consistency across a system. Generating one image is easy. Building an identity that survives fifty applications is a different discipline.
Defending a decision. Sitting with a client, explaining why this choice is right, and either changing their mind or changing yours.
And there is a craft detail anyone who has held a camera knows: generated output looks dead for specific, nameable reasons, which are in a cinematographer's eye.
The Arabic angle
Here the gap is wider. Models render the Arab world through stereotypes, fail at Arabic text inside images, and blur the region's distinct visual dialects. That is the argument in why AI draws the Arab world wrong.
This is not a complaint. It is an advantage: knowing the difference between a Najdi majlis and an Emirati one is something the model does not have and a distant competitor cannot fake.
What to actually do
Stop pricing execution by the hour. Price the outcome or the effect.
Master direction rather than description. The difference is everything, and it is in stop describing, start directing.
Build a reputation in something specific. The model is general, and general is now free. Specific is not.
In closing
Design will not vanish, but the market for execution is shrinking fast and the market for judgement is widening. The designers hurt most will be those who thought they were selling taste while selling hours.
Try this today: look at your last five delivered projects and ask, for each: was I paid for a decision or for a file?
Common questions
- Will AI replace designers?
- No, but it has replaced much of the execution: first concepts, variations, resizing. Visual judgement, understanding the problem and building a consistent system remain human work.
- Which skill should a designer develop now?
- Visual judgement and direction, meaning the ability to look at thirty outputs, pick the right one, and justify the pick. The system generates but does not discriminate.
- Do generative tools work for Arabic content?
- Partly. Arabic text inside images is still weak and the region's visual representation is stereotyped. Knowing the cultural distinctions is a genuine advantage here.
- How should I price after this shift?
- Move away from hourly pricing on execution and price the outcome or the effect, because execution cost collapsed while the value of the right decision rose.
No comments yet
Leave a comment