Recently I interviewed BotWiki founder Stefan Bohacek on the distinction between art made by a human and art made by a machine. “If you think art is these deep thoughts expressed by a human, then of course what bots make isn’t art”, he said. “If you think art is anything that looks good, then bots make art”.
Thinking more on the issue, I realized that a bot’s output is just the randomized result of human input. Even advanced bots with neural networks either learn from human input or learn from other bots that were programmed by humans. In the end, there’s no distinction between art created by bots and humans because humans are the ones that set boundaries for the bot and say what it can and can’t generate. If you can blame the programmers of abusive bots because they weren’t programmed with precautions, you also have to admit that bots are creating art and the humans are just as responsible for that art as they would be for painting a picture.
This being said, generative digital art is clearly considered to be worth less than the classic human-created works we’re all familiar with.
When a person paints a picture it’s the effect of random environmental effects. What they’re feeling that day, what happened, what’s on their mind. The same way that bots randomly generate art inside a preset framework, humans do the exact same thing. It’s just as unpredictable to guess what an artist might be feeling on any given day as it is to guess which exact values and parameters a bot will choose before it triggers off a piece of generative art.
The obvious argument is that art is some deep gaze into the human condition and that bots can’t possibly grasp that. Of course, bots aren’t advanced enough to feel emotions, but isn’t it the emotional impact of the art on the audience that counts?
In my experience, bots can create emotional art. Poem.exe‘s generative haikus, and the dreamy mountain ranges of soft landscapes in particular.
a flash of lightning
how about this
my bottle, empty
in the snowy road
— poem.exe (@poem_exe) May 29, 2017
— soft landscapes (@softlandscapes) May 21, 2017
When bots break into reality
Like most internet creations, bots usually stay confined tightly to the digital realm. The physical world and the digital datascape are two very separate things. However, there are some cases where bots breached the boundaries.
Random Darknet Shopper, programmed to purchase $100 of items from the dark net once per week, mailed the items automatically to an art exhibit where they were put on display. The project is an exploration of the dark net, shown through a wide and randomly selected sample of the available goods. Included were a .pdf file explaining how to hack a Coke machine, a British Gas bill template, ecstasy pills, and a passport scan. The same artists, !Mediengruppe Bitnik, also built a real version of a randomly glitched building.
The Art of Bots, a London art exhibit that’s exactly what it sounds like, is just about the only documented example of bots in the traditional art scene.
And, finally, Bot Summit 2016 is a rare example of botmakers showing up to demonstrate their bots to an audience.
*Terrible video quality warning*
With so few cases of bots existing in the same physicality as humans, it’s hard to believe the traditional art world is ready to accept bot art as a valid medium.
Twitter bots especially; they’re displayed on a platform that’s designed for the very opposite of deep analysis. Users are in the habit of quickly scrolling through the timeline, not deeply engaging with the content.
For that reason and the reasons I’ve mentioned already, bot art seems set to remain a kind of secret society, clustered around a few niche web communities and hashtags, leaving the hallowed spaces of museums and galleries for art that, for some reason, is treated with more sincerity and respect.
My 1hr interview with Stefan about the philosophical side of bots will be published soon. Follow Secret Cave on Twitter to get notified when it’s out