We wanted to run through a brief history of our own logo creation and how we looked to AI image generators as co-collaborators along the way – AI generators and logos, what’s the potential?

AI generators and logos – design concepts and initial ideas

First, we looked to AI image generators to throw a few ideas around. Part of our text prompts that we used included initial visions and vague concepts for our icons and visual elements, and we took those and added text like “black and white clip art” to see what came out.

The initial round of concepts were all over the map. Lots of unintelligible symbols, imagery that was too chaotic or unrelated, and some other rejected ideas. We did find a few images using AI that had certain elements we had not thought about yet, like different ways to think about nodes in an AI’s decision tree.

AI generators and logos – running human-made versions through AI software

Next, we took our human-made versions and ran them through AI software to find unique spins on our designs. This led to a number of new ideas, a couple of which were used by our human team in the final revisions. As far as AI generators and logos go – this was proving to be a valuable experience, and something that was important to us to take part in. After all, we are a machine learning art company, so why not collaborate with AI on our very logo?

AI logos – the story behind our own logo

Our logo has a lot of nuanced meaning baked in. We have a robot with a “third eye” – both as consumer and producer of art. This speaks to the Divine Spark of creativity for us – how much of this art is machine as artist, and is that even truly possible?

We’ve seen machines automate manual tasks for us throughout history, like assembling products – but what about unique creativity? Is it real art? Whatever it is, we understand that the software is both inputting images as well as outputting them.

It is trained on weighted and captioned images in libraries we give it, and it generates imagery based on that training and library along with our commands. For this reason, we still run with the idea of a third eye, and explore further when we begin adding elements of what appears to be radiating beams connected to the robot’s third eye.

At first glance, it may appear that the robot is taking a picture – a callback to flash cameras. We see what looks like light to be emanating from the robot’s third eye, which leaves one to assume it is inputting surroundings just like a camera would.

We also see these beams as projections and different iterated output the robot is distributing to the world – different variations of the same image idea.

Further, we see the lines from the eye along with the other two eyes as an obfuscated, abstracted visual representation of machine learning nodes and pathways between nodes.

Next, we see another point of input and output – the robot’s mouth. This represents other AI content, like AI generated text. Something else to think about here is that the old saying “you are what you eat” maybe applies to AI generators and logos (as well as other images). We cannot expect better output than the accuracy, quality, and richness of captioned libraries that the software is trained on. The jaw of the robot is a flat base, representing the base underlying technologies – with a straight blank space completely disconnecting the jaw from the head, to represent how output, while trained on specific images, can be completely unique from any single source image or collection of source images.

Finally, we give our robot ears to represent AI music and speech generators.

Some AI variations of our logo, isometric style!

Please enjoy a small collection of AI images where our logo has been run through AI image software using isometric styles.