This is a continuation of of maze design experiments. Read about it in post #1, One Maze - Many designs, then The first 5 experiments and so on, or just enjoy the ride.
The original maze:
I used Bing text to image for all of these creations. This is similar to my use of Google Gemini Remix a few weeks ago in the 6th part of this series. I was using Bing to test how their image generator does at making maze art (Making mazes with AI: Bing AI Image Generator - Has it improved over time ?) and I saw templates at the bottom of the page, so I decided to use them here. Basically I uploaded my starting maze and hit the template button and saw what was generated. Here is what I found:
The first maze is the cartoon sticker template.
The second maze is the coloring page.
The third maze the sketch template.
The fourth maze uses the Ukiyo-e print template.
The fifth maze I used the regular text to image and wrote “Make this maze more interesting”. This is what Bing came up with.
The 6th and 7th were so far off, I included them but do not call them mazes. I used the Minecraft Me and Make Me a toy templates. Both fun but too far off in this context.
This was a flop, but that is why I text AI image generators. The maze looks stayed, but the changes made them non-solvable mazes. This does not surprise me based on What I learned using AI to make maze art.
