Did a robot write this?
If you ever find yourself chatting with a writer at a party, likely as they stand awkwardly in the corner of the room avoiding all eye contact, try not to mention the fact that robots are capable of writing now and that they’ll probably take over all the writing jobs soon.
Not because robo-writers scare us. We all know the only thing that scares a writer is a deadline.
It’s just that talking about AI, machine learning, and robots forces us to admit something that we normally like to keep quiet — writing is really hard and not always fun, and some of us will take all the help we can get.
Friendly machines
Writing with the help of machines isn’t a new thing, nor is it inherently frightening.
I’ve come to really appreciate the software Grammarly, which uses AI to spot my wordy phrases, vague language, passive voice, and other stylistic issues. And don’t even get me started on the merits of using Word’s spell check. While I have a love/hate relationship with Google autocomplete, I can admit that sometimes it’s a huge time saver.
Writing is just like any job where the automation of repetitive tasks can actually make things better.
“The Washington Post’s robot reporter has published 850 articles in the past year” is a clickbaity headline that a robot probably wrote. If you read the article (which apparently many people aren’t doing these days, yet another thing to avoid telling a writer at a party), you’d learn that Heliograf, the Washington Post’s AI, is mostly writing articles that wouldn’t have been assigned to a person anyway.
By using it, the Post managed to free up some time for reporters so they could focus less on tweets and headlines and more on, you know, actual reporting.
I recently read a quote from Google UX writer Roxanne Pinto that sums up the current machine-writer relationship quite well:
“Since we (writers and designers) are the keepers of the stories, the only jobs that AI can take from us are the ones we want to give away. You can teach a machine what writing is, but it can’t write a UI, description, or article on its own — it needs your ideas first.”
I appreciate her saying that we are still somewhat in control of the situation, but I’ll be honest, I can’t seem to get those last five words out of my head: it needs your ideas first.
Whose idea is it anyway?
For creatives of any kind, our ideas are all we have. Forgive me if I’m a bit possessive of them. I’m perfectly happy to let a machine take the reins for a headline, task-oriented conversation, or even an article if it’s something boring that I don’t want to write about (Take that, robot!).
I have an issue with machines encroaching on the world of fiction, probably because I’ve read enough sci-fi to know that it can’t possibly end well if they get involved. Take the short story “The Great Automatic Grammatizator.” It’s about a man who creates a machine that can write entire novels in mere minutes. The machine eventually becomes so successful that it takes over the whole industry, leaving the majority of writers out of work, without a paycheque, and even more miserable than usual.
Roald Dahl wrote that story in 1953. I read it when I was a kid and thought it was interesting (though not nearly as interesting as that other thing he wrote about the chocolate). I knew, however, that there was no way I’d let some machine take away the most amazing thing in the world to me, which was, and still is, coming up with ideas, putting them down on paper, and having other people read them.
Today, when I see headlines like “Japanese AI Writes a Novel, Nearly Wins Literary Award” or “AI Is Beginning to Assist Novelists,” I think of Roald Dahl all those years ago writing a story, which turned out to be predictive fiction, and I realize that not everyone thinks machine novelists would be such a bad idea. This is clearly the beginning of a new and very complicated relationship.
No threat to humankind?
When you dig a little deeper, you discover that the Japanese AI-writing novelist was actually a team of real human people feeding select words, sentences, and parameters to a program, which then “wrote” the novel.
And Robin Sloan, the novelist being assisted by AI, is using a machine learning system that he created, which combines a recurrent neural network with a database of old science-fiction magazines. Sloan uses machine learning to generate auto-completes for his sentences and to come up with purposefully stiff, alien-like dialogue for a character in his book.
Neither of these scenarios sounds like a real threat to writers. AI still needs your ideas before it can do anything. And while I don’t think there are a ton of people working on giving robots the ability to come up with intricate plots and memorable characters on their own, I recognize that writing is a skill. It’s a craft. And if I learned to become a better writer through years of trial and error, why couldn’t an algorithm do the same?
Will we ever get to the point when AI can write something indistinguishable from something a human wrote? Probably. But maybe the more important question is, why would we want to?
