In my process of creating my own story in the world I have created last week, it was a complicated process. It was difficult making sure all the details added up for what I was going for. My plan for this story was that I wanted a man to attempt to return his AI automaton. Since this was made in the world of a utopia, I had to find a way to cause conflict between two characters. It also had to relate to AI. So I created a story of an old man named Lucas being stuck in his ways and not realizing that AI is not meant to replace him but to help him live his life.
Here are some of the challenges I had maintaining consistency. One of my issues was that I had to keep on regenerating images for the storyboards until I got the image that matched up with what I had imagined in my head as best as I could. My other challenge was that I had to retrieve certain bits of information used in previous storyboards so that the next one lines up and looks consistent like making sure this image still takes place on a street and will the cane appear or not. I tried using the IMAX film format on my storyboard but the AI told me that DALL-E cannot replicate this style so I just chose to make my images 1920×1080. The GPT still had issues being accurate with that resolution. My last challenge was that I now had to write the GPT to never include a cane for Lucas because he keeps being generated with one.
Now I will explain how I used the AI tools to generate and refine my world and its visual representation. When I am not using close up’s, medium or long shots, I also used over-the-shoulder and Bird’s Eye view. These are definitely difficult for Chat GPT to understand since it didn’t do exactly what I asked it to.
Here is where you can find the presentation.