We were at the hotel bar, our drinks forgotten as I thumbed a prompt into the chatbot’s mobile window. The prompt was something I’d arrived at like an improv comic—gimme an object, gimme an activity—though my audience had been only Jes, and we had already exhausted the novelty of asking the chatbot for tongue twisters and socialist manifestos in iambic pentameter. This time, we’d settled on in a paragraph compose a feminist analysis of Banksy’s Flower Thrower. My finger hovered over SEND.

I looked at Jes. “Want to see me lose my job in real time?” I said, and tapped.

We watched the black box of the cursor blink, blink, blink. Then at last it started to move, flickering across the blank white of the screen, pushing out words behind it.

“Unreal,” breathed Jes.

The chatbot we were huddled over, officially called ChatGPT, is an artificial intelligence developed by OpenAI, a research lab based out of Silicon Valley. It’s been making waves online since its free-to-use research preview debuted on November 30. Trained on a massive amount of data to replicate human speech, ChatGPT responds to user-generated prompts—some of them truly hair-brained—with organized, grammatical, and original sentences. And while the chatbot struggles occasionally with factual error, the sheer variety of these responses are stunning. With little more than a sentence to go from in some cases, ChatGPT can compose sonnets, troubleshoot code, draft emails in languages the user doesn’t know, answer questions about nuclear fission, summarize articles, and unpack complex philosophies.

As Jes and I learned at the bar, it can also compose serviceable feminist criticism, of the kind I’d expect to receive from my first-year undergrads:

Banksy’s Flower Thrower is a thought-provoking piece that invites feminist analysis and interpretation. At first glance, the image of a person throwing a bouquet of flowers seems benign and even joyful. However, upon closer inspection, the figure’s obscured facial features and riot gear suggest a deeper, more complex message. The anonymity of the figure could be seen as a commentary on the ways in which women’s voices and identities are often erased or marginalized in mainstream society. The riot gear, on the other hand, could be interpreted as a symbol of resistance and rebellion against the patriarchy and its oppressive systems and structures. Overall, Flower Thrower serves as a reminder of the ongoing struggle for gender equality and the importance of feminist activism in challenging and disrupting the status quo.

Not bad! Humdrum opening and a little surface-level but points for clarity and specificity. Writing in The Atlantic, Stephen Marche pegged the chatbot’s prose at a B+ on an undergraduate scale. “The passage,” he notes, “reads like filler, but so do most student essays.”

Marche, whose own essay is titled “The College Essay Is Dead,” is not alone in drawing a straight line between the chatbot and education. Although ChatGPT promises to reshape work and play outside the classroom, its astonishing capacity to automate—or anyway to simulate—writing and critical thinking has prompted a flurry of articles, thinkpieces, and social media posts. Much of the discourse, panicked, breaks in predictable but not unwarranted directions. ChatGPT will enable cheating. Teachers will have to rethink whole curricula. The college essay is dead, high school English is dead, tried-and-true justifications for an education in the humanities are dead.

Other approaches have been less dire. Finger-wagging at alarmists for playing up the classroom as an antagonistic space—the overworked instructor vs. the cynical, shirking student—members of this latter camp acknowledge the sea change ChatGPT means for education as well as the adaptations it will necessitate. But they also argue that AI chatbots can, potentially, present opportunities: to reassess the goals of writing, to educate about the bias and ethical limits of AI, to incorporate AI into the writing process and instruction. Provided teachers are adequately supported, ChatGPT means change, not an end. As Beth McMurtrie, a senior writer for The Chronicle of Higher Education, puts it: “[w]hether AI chatbots become a faculty nightmare or just another teaching tool may ultimately come down to this: Not the state of the technology, but whether professors”—or teachers in general—“are allowed the time to create meaningful work for their students.” (The Chronicle paywalled this article; message me if you’d like a pdf.)

And that, it seems to me, is the kicker. AI like ChatGPT will shape education going forward, whether teachers like it or not. So rather than treating it as hostile to the goals of education—an obstacle to be managed, solved, or mitigated—the ideal solution is to provide instructors the time, training, and resources to adapt. But time, training, and resources are a perversely tall order these days. In the face of chronic overwork, the casualization of the educational labor force, and mounting hostility to public education, it’s difficult to imagine the material conditions of instructors improving much in the near future. Far easier is it to envision a situation where instructors find themselves even more exhausted, scrambling to rejigger assignment sequences and learning goals in a desperate effort to outpace ChatGPT’s blink-blink-blinking cursor.

Put differently, the problem with ChatGPT is not its astonishing novelty. Not really. The problem instead is how it exposes and exacerbates already existing contradictions in the profession—contradictions that even a fancy new tech, one that, with equal aplomb, drafts workplace emails and composes halfway decent sestinas about Martin Freeman, can never hope to resolve.

Submit a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

post calvin direct

Get new posts from Ben DeVries delivered straight to your inbox.

the post calvin