I think there’s often a misunderstanding about how AI “learns.” It doesn’t copy and paste bits of existing art. During training, the model is shown millions of examples to understand patterns—like teaching a child what blue is or what an apple looks like.
Once training is finished, it generates new images from what it has learned, not by pulling anything from the original sources.
For me, working with AI isn’t that different from using commercial-use resources: it’s a starting point to create something uniquely mine. And for me it’s certainly not a shortcut around creativity. Shaping a prompt to get the result I’m after takes plenty of thought and experimentation, and then there’s the final stage—pulling the different papers and elements together so the kit feels like one coherent story. That’s my own process; other designers may approach it differently.
Like any tool, AI can be used well or poorly. It can open amazing creative doors, but it can also make copying someone else’s design easier than before. That risk is real, and it’s one reason I stay very deliberate about how I work and what I release.
As Ona said, the real power is that AI can help artists and designers translate the ideas in their heads more easily than starting completely from scratch in Photoshop. I also spend a lot of time after generation refining in Photoshop—checking anatomy, perspective, lighting—until everything fits my style and quality standards.
I understand why the topic feels tricky, but to me it’s simply another tool in the studio. It isn’t going away, and it keeps improving.
In the end, what matters is the creativity and care you bring to the process and the finished work your customers see.