But Is It Art?

But Is It Art? Artificial intelligence, as I’ve written in the past, is an exciting and powerful means of computers completing tasks which would previously have been impossible or too time consuming to be justifiable. Another more recently popular use for AI is the generation of images from simple text prompts. Say you’d like an AI’s idea of how a cat looks, you’d program it to analyze a set of photos of actual cats, then return what it thinks is the most accurate rendition based on that data. In the early days, this produced results which only had merit for comedy value, but as the technology has matured it’s now raising questions about whether we can go as far as considering it art. Is this a shining example of our progress with machine learning, or a damning indicator of how loosely we label something as artwork?

First, some quick clarification for those who missed my previous pieces on AI. Artificial Intelligence doesn’t (yet…) function like the scary, world ending, Skynet-esque computer baddies which movies and other media has led us to accept. It’s mostly used for recognizing patterns and is deployed in things like obstacle avoidance, marketing and optimizing smart devices. Anyone with a modern iPhone can witness it in action, simply by heading to your photos app and using the search function. Queries such as “boat” or “trees” will return images which an AI has analyzed and tagged as probably containing those items. A very humanlike task, performed to a surprising level of accuracy by a machine.

Now using that knowledge, it’s not too much of a leap to imagine we’d be able to ask that same machine what it thought a boat or a tree might look like. It could then generate its best guess for us to view and decide how close it was to our own interpretation. What gets really interesting is when we ask these AIs to show us images of abstract ideas, such as joy, fear, moroseness or desire. Here’s where we begin to blur the line between what’s simply an image generated by a machine programmed to trawl photos and return a result, and something resembling a piece of art created with depth and consideration for the subject matter. Art created by a hippo bumping its nose into a canvas sells for hundreds of dollars, is this really any less valid?

This question has recently been polarizing opinions, with a piece generated by AI having won first prize at Colorado’s State Fair, in the “Digital Arts” category. On one side, people have defended its creator Jason Allen’s win, given he set out to make a statement about using artificial intelligence in a competitive manner. It wasn’t without work, too; he generated hundreds of images, tweaking his methodology each time before settling on the final imagery. On the other however, there’s an argument being made centered around how if he didn’t draw or render the pieces from his own imagination, does it qualify for the ribbon, or indeed as art at all? As it stands today, creations such as this exist in a grey area until they become more mainstream and accepted.

Art is by its very nature, subjective. As the future becomes the present and we welcome the acceptance of achieving things by means we aren’t used to, this is perhaps an issue which solves itself. AI generated artwork probably stands alone as a category right now, the question maybe is how long until it’s normal to peruse a gallery and see it hanging side by side with pieces penned by conventional, human creators?