The Iconic Vision Of Google AutoDraw
When Irregular Times began, back in the mid 1990s, the Internet in its popular form was brand new. We started out printing on paper, although we used computers to format the type and the primitive graphics we employed at the time. Now, over 20 years later, computers aren’t just formatting and transmitting text, they are creating it. Computer programs are being used to automatically write some forms of news articles.
It has long been said that repetitive tasks will soon be taken over by machines, but now people are beginning to predict that creative work, such as writing and the production of visual art, will be shortly be automated as well.
Profitability for big corporations should theoretically skyrocket when they can replace human workers with computers that never need to sleep and have no family obligations. Of course, that profitability won’t benefit the population at large. A handful of investors stand to make a killing, while the rest of us will merely get slaughtered. The history of corporations that promise to share their bounty is a history of broken promises.
Back in 2012, Christopher Steiner of the Wall Street Journal wrote, “Creative types tend to think of themselves as doing work that is beyond the reach of automation. Computers can’t parse nuance, the thinking goes, or summon the imaginative powers that are required of writers, artists, technological innovators and policy-makers. As it turns out, however, this flattering assumption is mistaken. Computers can be creative after all. The more we understand about creativity, the more we are able to distill it into the language of algorithms—the ‘brains’ behind computer programs.”
Was Steiner right? Will creativity show on be reduced to an algorithm for computers to pump out, putting artists out of work? I decided to see for myself what artificial intelligence is currently able to make of art. I went to the heart of machine learning: Google.
As a part of the corporate movement to asset the growing relevance of artificial intelligence, Dan Motzenbecker and Kyle Phillips at Google Creative Labs have created AutoDraw, a program that claims to sit at “the crossroads of art and technology, created by artists and creative coders”.
“AutoDraw’s suggestion tool uses the same technology used in QuickDraw, to guess what you’re trying to draw. Right now, it can guess hundreds of drawings and we look forward to adding more over time,” they explain. Working with AutoDraw is supposed to be like working with AutoCorrect. You go to the AutoDraw web site, try to draw something on the screen, and then allow the AutoDraw algorithm to choose a sleek, smooth drawing to replace it.
What actually happened when I used AutoDraw was quite different.
I drew a bed.
Google Autodraw concluded that I drew a shark.
I drew a bottle of Coca-Cola, even putting the word “Coke” on the front.
Google concluded that I drew an owl.
I drew an octopus.
Google Autodraw concluded that I drew a person doing yoga.
I drew a person climbing a mountain, with some mountain goats.
Google Autodraw concluded that I drew the Eiffel Tower.
I drew an artist painting some flowers.
Google Autodraw once again concluded that I drew the Eiffel Tower.
I drew Krampus.
Google Autodraw concluded that I drew a tree.
AutoDraw couldn’t grasp what I was trying to draw. All it could do was reduce my creations to their simplest geometrical essence, and compare those reductions to the icons already in its collection.
I am by no means a skilled artist, but any human being could look at my drawings and see what I was trying to sketch. AutoDraw’s so called artificial intelligence is incapable of doing that. It can’t even achieve the simple task of seeing a simple drawing in a controlled online environment.
If artificial intelligence from the global center of machine learning can’t interpret visual art, it can’t create it. Yes, there are wonderful digital tools that can be used by people to create and enhance visual art, but the algorithms that make up these tools aren’t themselves creative. People need to use them to achieve a creative purpose.
Creativity, after all, isn’t a matter of technical expertise or compliance with a well-designed formula. Creativity begins with with purpose, and artificial intelligence does not work out of its own sense of purpose.
Some day, the people at Google Creative Labs may figure out how to craft an algorithm that can recognize the difference between a bed and a shark. Even then, the company’s computers won’t be able to understand what the difference means.
Art isn’t just about making things pretty. It isn’t about making lines straight. It’s about making inspired expressions about what matters.
As more of the world is sorted and ranked according to Google’s systems of artificial intelligence, the need for human artists to help us understand what it all means. Subjective creativity may be the only thing that can save us from the objectification of human experience.