Dall-E & Me

2023

Note: This is a heavily condensed version of my thesis. If you would like to see my entire process book, which further details my research methods and how I got to my solutions, click here.

Part I: From Artificial to Corporeal

When an artificial intelligence like Dall-E can bust out stunning furniture designs like these in minutes, how do designers fit into the industry? What is our role? How can we harness this technology, and what does the future look like for us? These are all questions I had tumbling through my head as I researched my thesis.

Moreover, what is the "new" design process like? What does it take to bring one of these generations into being?

Naturally, as a product designer my first instinct was to get on creating some 3D models. I was curious about the 3D modelling AI that OpenAI was releasing though - Point E. Like Dall E, this model creates 3D meshes from text prompts. I was also able to find a way to take my Dall E images and use them as prompts, but it was less than successful. The image meshes were closer to broken shards than full objects.

Overall, Point E isn't as sophisticated as Dall E yet. But one day, it'll be just as easy to create detailed 3D models as it is images from Dall E.

So, back to current methods we went. I picked some chairs from the 100 total Knoll-esque chairs I generated, and threw them together in Fusion. Afterwards, I printed one of the models in SLA resin, and used those measurements to reverse engineer the full scale prototype.

Full scale foam chair versus the 3D print. The legs are thicker on the full scale due to the type of foam I used to carve it out. I used a mortise and tenon technique to secure everything together.

Part II: Beyond the Algorithm

I started thinking about how designers would learn to use AI tools, so I threw a storyboard together. I wasn't sure if I liked this process though - something about it felt un-intuitive. Perhaps because this is more a "solution for extroverts" and I myself am a quintessential introvert.

So instead of a studio service design, I decided to create a different app interface altogether. Something that you can use as a design tool, and have it teach you how to refine your prompts at the same time.

Here is my solution - Chemist. An all in one intuitive AI prompt trainer and design tool.

Part III: An Unexpected Ending

I was so shocked, and pleasantly surprised to be the recipient of the OCAD University Best Thesis in Industrial Design medal for my grad year. This project had a lot of challenges thrown in the way, a lot of people misunderstanding my goals. But I kept to it!

This project is not a sign of support or opposition to artificial intelligence. Merely a window into what we can expect.

I am forever flattered to have received my medal.

Thank you.