Bricks & Brands
What if today's popular brands had Lego sets?
Our Bricks & Brands project combines two of my favorite things: Lego and brands. We asked AI to visualize brands' imaginary Lego sets. Of course, we provided some guidance to the AI during the production of the project, but we tried to keep that guidance to a minimum. We wanted to leave room for randomness and surprise about the brands. We only made small corrections (thanks to Adobe Photoshop's Generative Fill function) and tried to leave the resulting flaws as untouched as possible, because we believe that flaws can be just as revealing as beautiful details. (For example, you may see uncorrected errors such as "Nego" logotype instead of "Lego" in the Nescafe image, or some broken logotypes.)
If we look closely at this AI-powered project, we learn about the limitations of brands, their identities and guidelines, and how much of the information they convey is understood by artificial intelligence/machine learning.
Many years ago, we created the "3.5 Inch Poster Series" and showed how many floppy disks it took to install today's software. As you can imagine, some of them required thousands of disks. In other words, getting rid of the limitations of floppy disks for installing our software has expanded our capabilities beyond our wildest dreams. I think it would be a challenge to produce such a project without artificial intelligence tools. So while we are exploring our limits, we are actually traveling to new horizons. As artificial intelligence continues to learn from our creations, we, in turn, will have much to learn from it in many ways. As Antoine de Saint-Exupéry said, "The important thing is what you do not see."
P.S. At the end of the images you can also find more information, such as how many words are in the prompts we produced in the project, how many images we produced together with the ones we eliminated. The images we eliminated were distorted, erroneous images created due to the current capacity of the AI. Maybe we can show them in another project. Because there are some really interesting and funny absurdities.
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Facts & Numbers
148 images were selected for 37 brands, but 2873 images were produced for a total of 65 brands,
in 5 days with the help of artificial intelligence.
Nearly 10,000 words of prompts were written for the AI to generate these images.
-
The most difficult brands to produce images for were Shell, Heineken and Nestle,
where the AI produced dozens of incorrect images. The easiest were Ikea, Rolex and Apple, where the AI generated images with almost no errors.
-
We were also impressed by dozens of details that were not in our prompts but were described by the AI, such as burgers at McDonald's, shipping boxes at Amazon, parking spaces at Ikea, speakers at Spotify, coffee cups at Starbucks and so on.
-
Perhaps due to the database within AI's reach, although some logos have been updated, AI has preferred to use the old versions, for example, in the Master Card visual, it chose the old Frutiger version of its logo instead of the FF Mark version, which was updated in 2016.
-
AI was also able to correctly use different logos for different time periods, as can be seen in the Shell variations; in one image reflecting the '50s and '60s, AI chose a logo with uppercase letters, while in modern images it chose lowercase letters.
-
Legal Disclaimer: Copyright of all trademarks, brands, logotypes, featured background images
and photographs are the intellectual property of their respective owners.
This project is simply a design exercise to explore how brands are perceived by AI.
-
All of the visuals in the project were created with AI using written prompts.
Minor corrections were made to a few images using the Adobe Generative Fill feature.
-
