A few weeks ago I asked an AI to generate twenty variations of a book cover. In four minutes I had twenty designs. All technically clean. None of them were good.
None of them looked real. The proportions were right, the colors worked. But everything was synthetic. Something was missing everywhere that you couldn’t pin to a single element. The sense that someone had thought about it. That someone had wrestled with the material and arrived at a decision.
But that wasn’t the real problem. The real problem was what those four minutes did to me.
Before the four minutes I had images in my head. Ideas of what the cover could look like. A sense of direction. I was looking forward to the design process.
After the four minutes that was gone. All of it. The twenty synthetic suggestions had overwritten my own images. Four minutes, no result and my own creativity destroyed.
I would have expected a tool to support the process. This one ended it.
Designing a cover doesn’t mean producing an image. You engage with the content, ask yourself what the book wants to say. You make drafts that don’t work and understand why. At some point you have a feeling you can’t yet name and you work on the form until the feeling is in it.
That takes time. Sometimes days. And that’s exactly why something emerges in the end that you can feel.
AI skips this process. It delivers the result without the journey, and so without a story.
Maybe it just needs development time. Maybe in two years AI will deliver covers that look real. That may be. But even then the journey will be missing. And without the journey the result is missing a relevant part of its story.
I use AI for research and for structures that I then rework. That’s useful. But for things that should have something of me in them, I don’t use it anymore. Not because the results are bad. But because after four minutes I was sitting in front of twenty images and none of them were mine.