AI and the Democratization of Design

What happens when AI not only designs, but also takes over the thinking? A look at the role of designers and how to evaluate AI-generated designs.
Collage aus retro geräten

Basically, anyone can now do what I do.

A lot has changed. AI has led to an unprecedented democratization of design. A bit like when mobile phones first got cameras: suddenly, everyone was a photography expert.

At its best, AI support can lead to better design, even on small budgets, because many repetitive tasks are eliminated, saving a significant amount of time. Anyone can visualize and communicate their own ideas quickly, easily, and professionally, regardless of their skill set.

So why do we still need designers?

Have I become an annoying extra step between clients, developers, and the finished result? If that were really the case, I would have opened my cinnamon roll bakery a long time ago. But the visual end result is only the tip of the iceberg. Many steps in the design process can’t simply be skipped, even with the support of AI.

„AI makes building feel like thinking.“

Thinking Is No Longer Necessary

This quote by Robin Bigio has stuck with me:

“AI makes building feel like thinking.” – Link to the article by Robert Bigio.

I can confirm this from my own experience. With the user-friendly chat interface, the AI’s constant positive feedback, and the speed at which things happen, you gradually start switching off your own thinking. But if we leave all the thinking to AI, who decides what is good and what is bad?

What happens when you hand over an interface design, a website redesign, or a logo redesign entirely to AI? No research, no strategy, and ideally an instant, perfectly polished result?

Potential Problems:

  • You get ten different versions (in every imaginable color), but the target audience has been ignored.
  • The person behind the machine lacks the ability to reflect on the final result and question it critically.
  • You choose the design that feels safe, not the one that is actually needed. Very few people can immediately accept good design. It takes time, and it takes someone who is willing to stand behind bold decisions.
  • AI-generated design can encourage content overload. The assumption becomes: more is better, more is safer.

That’s why designers often act primarily as decision-makers. They don’t just decide what should be added, but above all what can be removed.

Or, to put it in the words of Dieter Rams’ tenth principle of good design: Good design is as little design as possible.

How Can I Evaluate AI-Generated Design as a Non-Designer?

  1. Focus on the actual goal (define it BEFOREHAND)

    Does the design achieve the product’s goal — regardless of what the AI claims? A design can look modern, professional, and impressive while still solving the wrong problem.

  2. Ask for explanations

    If the design comes from another company, question the decisions and ask for clear, understandable reasoning. This is where you can see the difference between people who created a design together with AI and people who simply switched off their brains.

  3. Ignore the design for a moment

    AI-generated designs often look especially polished, modern, and finished. That can make them harder to judge. Look at the design in black and white. Ignore colors, images, and effects for a moment. Focus on content, hierarchy, and structure.

    Do the decisions still make sense when the design no longer looks as attractive?

  4. Get a critical eye involved

    Ask someone with design experience to take a look — someone who has developed a “critical eye” over the years.

Conclusion: So why get a second opinion after all? Not because designers are always right. But years of dealing with design decisions build not only knowledge, but also intuition. And since bad design can become very expensive once it is out in the world, a second look certainly doesn’t hurt.

I wish you success in questioning things critically. And if you need support, feel free to get in touch with us!

This text was started by hand, developed independently, and linguistically refined with the help of AI.