Automatic writing*

A few posts back, I mentioned I would weigh in on artificial intelligence prose generators–“bots”–specifically the much-reported-upon ChatGPT. The media coverage has included everything: hand-wringing, speculations on the extinction of critical thought, predictions about the death of the high school essay or the re-institutions of oral and handwritten exams, not to mention worries about spurious content and job loss among educated citizens, as well as wild enthusiasm for automating tedious writing tasks and excitement about where tech is taking us.

It’s not as new as most people think. AI has been providing customer-service responses and generating basic summary content for news-related websites for awhile; but OpenAI’s open-source platform, which is currently free and for use by anyone (not to mention educating itself as each user inputs prompts and questions) has so rapidly gained “tech adopters” that those of us who teach writing cannot ignore it completely. And we shouldn’t ignore it, but neither should we throw our hands up in surrender and predict the end of the art of writing as we know it.

A recent New York Times article reflects the kind of discourse taking place at the institution where I work. It’s fascinating to me to see how quickly the conversations have evolved in the usually slow-moving environment of academia. I find that at my college, my years of laboring with students who lack strong backgrounds in written expression or confidence in their writing have suddenly attracted the attention of full-time faculty members–they want to know how they can tell if students are using AI assistance to write essays (when said profs have no pedagogical experience in writing) and how to change the wording of their assignments to “fool” the programmed generators, among other pressing questions. These inquiries tend to come tinged with a sense of slippery-slope fallacy: does this mean academia will go to hell in a handbasket?

I refuse to send out a firm forecast, though my intuition says no; instead, academia, and society, will change.

And despite the daily-proven, scientifically-accurate, anecdotally-obvious FACT that change is normal and indeed necessary, most people (and their societies and institutions) fear change. Hence, the media and institutional brou-ha-ha.

Let’s face it, writing can be hard. There will always be people who do not want to do the work of writing from the soul, brain, heart, emotion, experience, dread, you-name-it. Painting is hard, too. But people who don’t want to practice and experiment with visual art can use paint by numbers, clip art, or AI. There will always be a few folks who learn to play an instrument for the joy of it and for the challenge of continually learning new approaches to the process of music making; the rest of us can be audiences, if we like. People who write because they can’t not write? They won’t use bots unless they want to experiment with them: make perverse use of the programs, play with them to see what the human’s skills can do in concert with algorithms, bits, bytes, and data. I know artists who are already collaging with AI-generated art to create new, human-mediated visuals.

I recognize the fear factor here, but I don’t buy into it because I am so curious about what will happen next. I’m interested to see how changes will occur, which changes will make a difference and which ones will just vanish, and whether pedagogy will develop toward, away from, or parallel to AI developments in numerous spheres–to name just three of numerous possibilities. Change is exciting, but it’s also hard. I can’t say I am as excited about adapting my fall semester syllabus to reflect whatever the university decides to do in light of ChatGPT, but since I’ll have to adapt to a new “learning management system” anyway, I may as well accept that “a change is gonna come.”


*For the definition of automatic writing, Wikipedia has a fairly complete page.

Photo by Startup Stock Photos on

Written by a human

Here’s a controversy for National Poetry Month–there are an amazing number of controversies surrounding poetry–which takes up the idea of whether a “machine” can write poetry. A good introduction is this CCR interview with Oscar Schwartz, who developed Botpoet as an experiment that is not so much about artificial intelligence as it is about what humans consider to be poetry. And perhaps about what language really is. If you follow the link to the site, you can participate in his research by playing “Bot or Not,” a game in which the player reads a series of poetic lines and then chooses between written by a human or not written by a human.

If you’ve read a great deal of classic and contemporary poetry, you may recognize some of the poems (I did); I suppose that is a way to cheat the system, since I have insider information. Nevertheless, I was wrong embarrassingly often. What, exactly, was I looking for in those words?

I think Schwartz is correct in his assessment of the more general population (though literary types may disagree with general assessments) when he says:

People generally seem to associate rhyming, “Romantic” poetry as being human. And they consider highly abstract, non-traditional poems to be of human provenance. Investigating as to why this might be the case is the project of my PhD.

He points out that written language is arbitrary and abstract, “an artificial medium” to begin with, and may have less to do with being human than we might like to think. Maybe the qualities that make a poem a poem are qualities that reside in the reader/interpreter rather than in the poet, another individual’s aesthetics or sense of what seems “creative.” That might be an unsettling thought for many writers, though it rather appeals to me.

Schwartz continues,

“So the results of Bot or Not, rather than telling us what human really is, is actually telling us that the category of the ‘human’ is an ideological, political space…The Bot or Not project works not because it tells us about computer software, but because it reveals things about what we assume to be human. It destabilizes the category of the human.”

As it turns out, the study of consciousness also tends that way–destabilizing our long-held category of what-a-human-is or what, if anything, differentiates us from other animals. Some interpreters of Zen philosophy suggest that Zen consists in finding balance within the inherent instability of the corporeal world. Or, perhaps, acceptance that humanness may be something we cannot categorize; the challenge then is to learn to flourish in a state of destabilization.

Let me sing the body electric…and the mind (possibly) electronic.

Walt Whitman in mid-life

Walt Whitman in mid-life