This story is over 5 years old.


What it Means to Be an 'Experimental Computer Poet'

A conversation with Allison Parrish, the artist behind your favorite Twitter bots.
Image: aristihouse

Allison Parrish loves words.

She loves putting words in order, in disorder, and in uncomfortable, erratic sequences. She loves words that don't make sense.

A Dada artist of sorts, Parrish calls herself an "experimental computer poet." She uses computers to find unexpected things about language beyond the bounds of human semantic constraints. Her conceptual poetry often takes the form of Twitter bots, like her beloved 2007 experiment, @everyword, which Tweeted each word in the English language over the course of seven years.


Parrish, who is 36, has been pouring over programming manuals since she was five, when her parents bought the family their first computer –a TRS-80 Color Computer 2. Now, she holds a master's degree from NYU's Interactive Telecommunications Program, where she currently teaches coding and linguistics. Parrish is the author of "Everyword: The Book" and a co-designer of the word game Rewordable, which was published by Penguin Random House in August of this year. Her newest book of poetry, "Articulations," comes out in January.

I spoke to the botmaker about making art that explores language and its endlessly imaginative mutations.

The following interview has been condensed and edited for clarity.

Motherboard: Why did you choose Twitter as a platform for your computer poetry when you started?
Parrish: When I first started doing stuff for Twitter it was 2007, back when it was the wave of 'There's this thing and it's bite-sized and you can follow it, it'll be in your feed sometimes but you don't necessarily have to pay it close attention.' It was this new frontier that was exciting to explore, but since the adoption of the algorithmically-curated timeline, using Twitter now feels like playing a slot machine–finely tuned to being addictive by doling out outrage and dopamine in measured amounts.

So when was the point your relationship to it changed?
I should have been disillusioned with Twitter as a platform earlier, but the 2016 US presidential election sealed it. They have consistently shown themselves to be unwilling or unable to stop abuse on the platform and their recent policy announcement about Trump's violent threats reads to me as though they just like being the bullhorn for a dictator. But I'm not saying that I'll never make another project for Twitter again. I've always held that bots can be an effective means of intervention and protest.


Everyword, your most well-known project, tweeted every word in the English language, in alphabetical order, to thousands of followers over the course of seven years. How did the idea for it come about?
Everyword was originally a project for an ITP class [at NYU] where we were talking about an artwork called "Every Icon" by John Simon. That project was a 32-by-32 grid of pixels that turned on and off every pixel of the grid in every possible permutation to eventually display every possible icon on the grid.

Everyword just started as a riff off that project and in response to the story people told about Twitter at that time which was, 'Oh, it's inane. People are just posting about their sandwiches or whatever'– a very different criticism than the one we have of it today. So I just thought of a way to criticize that by saying, 'Here's every possible word on Twitter, how does that fit in with your conception of the platform?'

The Everyword bot plays with words in a similar way as your Ephemerides bot, which takes a randomly selected image from NASA's database of outer planet probes and posts it to Twitter along with a computer-generated poem. How did this combination of space probes and generative poetry come about?
I'm very interested in space exploration and science fiction. The similarity I perceive between space probes and generative poetry programs is that both venture into inhospitable realms and send back telemetry telling us what they found. I think that computer-generated poetry is especially good at thinking about the ethical boundary between what is good exploration and what is exploitation instead.

And with creation and exploration through artificial intelligence also comes the complicated question of copyright and responsibility.
For me, the people that are making these algorithms are the ones that are ultimately responsible for them. And that responsibility works both ways. When I make a bot or a program that generates poetry–that is my means of generating poetry. It's just my poetry that I wrote in the same way that Jackson Pollock doesn't attribute his work to Jackson Pollock and Paintbrush.

I recently read a study from the Artificial Intelligence Lab at Rutgers about an algorithm taught to replicate existing painting styles to the extent that people couldn't distinguish them from the works created by humans. Is this what you're trying to do through your poetry?
I think that imitation is the most boring thing you can do with a computer. It is frustrating because a lot of the academic research on creativity in artificial intelligence right now is focused on how to make a computer do something that an artist normally does, to take jobs that previously required skilled knowledge or creativity and trying to do them through a machine instead. They want to throw art and poetry into the mix and I don't think it belongs in that same category.

Putting words in unusual sequences is a big theme in your work. Why are you so concerned with the concept of nonsense?
People talk about "sense" as though it were some inherent property of language that can be empirically tested—like you can feed it into a machine and get back a 100 percent accurate evaluation of whether it's sense or nonsense. What I'm trying to do as a poet is expand the kinds of interpretations that people can bring to language. Because computers have very few "built-in" ideas about how language works, they make you invent your own abstractions of language and are prone to doing unexpected and beautiful things when those abstractions go awry.