Select Page

While I was stuck in the elevator today because I accidentally pressed 6 instead of 4, then panic-pressed 2 for some reason, I started to think about fonts. Then the ride was over, and I forgot about fonts again because I had to start walking with an open cup of coffee, to which I had forgotten to pick up the handy lid that would have kept me from scalding my hands.

Safely at my desk, unscalded, I remembered my font thought from earlier. Russia released some terrifyingly Terminator-like “space robots” recently, which led me to recall the Skynet logo. How could anyone not have realized Cyberdyne Systems Corporation was going to spawn the end of the world?

Look at that evil font choice and sinister color scheme!

“Evil robot overlords” totally comes across in Skynet’s logo, but what about companies actually dabbling in AI and automation? Google, for one, is like the Labrador of logos: they want everyone to think they’re harmless, hiding behind their bright, candy colors and big, curvy font choices. Balloons! Circus! Clowns! Haven’t you seen Stephen King’s IT? Nothing ever ends well with clowns.

Eventually, my wanderings on the interwebs turned up this article about something called Font Map.

At first glance, Font Map sounds like a cool experiment.

Font Map interface

Reminds me of those huge swarms of starlings.

A designer at Ideo used machine learning algorithms to visualize the relationships between hundreds of fonts from Google Fonts. He tasked the algorithm with sorting 750 typefaces – serifs (the ones with the squiggles) on the left, sans-serifs (the ones without the squiggles) on the right, and everything in between. It’s interactive, informative, and fun (for type-crazed designers).

Then I read this part: “Perhaps one day, algorithms won’t be just categorizing fonts – they’ll be creating them too.”

That was about the time that the panic set in. I always thought the creative field was relatively safe from the coming automation revolution, but now I’m not so sure. Previous brushes with creative automation were strange and disturbing.

Google’s artificial neural network already at work giving us nightmares

Theoretically, we could ask an algorithm to meet certain parameters. We can show it what a “good” and “bad” font are. We can feed it information, then ask it to come up with a solution. This is a very real possibility. Once the AI knows how to make the fonts, how long before it starts drawing conclusions such as which fonts match others? How long before we just ask Google for a logo – and voila! There it is. Just add color. Export, or better yet, embed straight to your blog or website.

Where do we draw the line with automated design creation versus human?