Liberal Arts and Illiberal Artifacts

Essays are ubiquitous in education, where they serve many purposes, from demonstrating how students think to providing an opportunity for reflection. But they are also really easy for any Large Language Model to generate, as I think many educators know. In fact, many teachers have already seen obviously bad AI-produced essays and developed methods of identifying them. But none of those methods will work in the Fall . . . THERE IS NO WAY TO DETECT THE OUTPUT OF GPT-4 [emphasis original].  A couple rounds of prompting remove the ability of any detection system to identify AI writing. And, even worse, detectors have high false positive rates, accusing people (and especially non-native English speakers) of using AI when they are not. You cannot ask an AI to detect AI writing either – it will just make up an answer. Unless you are doing in-class assignments, there is no accurate way of detecting whether work is human-created.

My colleague avowed that, for his part, he would, henceforth, assign in-class essays only, lest he tempt his students to an unending orgy of indetectible, AI-enabled plagiarism.

In the course of that first exchange, I averred that I had never spent much concern on plagiarism in my tutorials and seminars, and saw no reason to suppose I would have the least difficulty recognizing a ChatGPT-product, should—unlikely event!—one come my way, since my essay assignments depend upon particular class conversations, and I require a first draft and accompanying draft-conference.  The instant response:  “Ask me for an essay and I’ll get you a first draft within minutes . . . [G]ive me a prompt.” And so—rashly!—I sent: “In Chapter 39 of Emma, Emma Woodhouse rejoices ‘that they [she and Mr. Knightley] had come to so good an understanding respecting the Eltons . . .’  Discuss why it is that Emma finds the understanding ‘so good’ and how it is that she and Knightley reach it.”  This prompt, dispatched at 1:14PM, was answered at 1:20PM by a “composition” of some 750 words, generated (my correspondent noted) “in under a minute.”

I shall spare the reader experience of the whole confection (though anyone may have it on request).  Four paragraphs—the introductory, the first substantive paragraph (and third in order), the fourth substantive (fifth in order), and the conclusion—will serve my (frankly polemical) purpose.

In Chapter 39 of Jane Austen’s celebrated novel, “Emma,” we find our protagonist Emma Woodhouse in a moment of delight. She is exultant that she and her friend and future love interest, Mr. Knightley, have reached a harmonious understanding concerning the Eltons, a couple around whom much of the novel’s social strife swirls. The concord they achieve is deemed by Emma as “so good” — a phrase worthy of examination within the context of the social landscape in which Emma operates and how such understanding has been arrived at . . .

Emma’s reference to the understanding as “so good” should not merely be taken as an expression of agreement with Mr. Knightley about their shared perception of the Eltons. There is a deeper implication in this simple phrase. It reveals Emma’s underlying aspiration for order, harmony, and status quo in her social world. The Eltons, with their often discordant behavior and ambitions, pose a disruptive threat to the equilibrium of Highbury society. By reaching an understanding with Mr. Knightley regarding the Eltons, Emma sees a way to reintegrate the perceived social misfits and maintain societal balance. It is the goodness of this balance, rather than the understanding itself, which Emma finds so pleasing . . . .

Emma, with her propensity to idealize relationships, initially favors the match between Harriet and Mr. Elton. It is her idealistic nature that blinds her to Mr. Elton’s social aspirations, leading to a sequence of unfortunate misunderstandings. Conversely, Mr. Knightley, with his incisive understanding of human nature and societal rules, perceives the implications of such a match and sees through Mr. Elton’s pretenses. His counsel acts as a tempering influence on Emma, nudging her to a more accurate comprehension of the Eltons and their societal role . . . .

In conclusion, the ‘good’ understanding Emma refers to is multifaceted. It signifies an alignment of perspectives, a testament to the power of discourse and respect, a mechanism for preserving social balance, and a stepping-stone in the evolution of Emma and Mr. Knightley’s relationship. It embodies the intricate dynamics of societal interactions and personal relationships, vividly illustrating Austen’s insight into the human condition. Through the lens of Hannah Arendt, this event may be seen as a testament to the ‘vita activa’, the active life, in which discourse, action, and power intersect to shape our collective existence.

We must, I think, agree that here the prognostications of “The Homework Apocalypse” are, like those legendary Soviet Five-year Plans, triumphantly over-fulfilled.  Who (or what on-line Turnitin-clone of a plagiarism checker) could distinguish these clotted banalities from the typical output of resolutely unthinking, net-surfing undergraduates, lost to education through credential-lust?  None at all, it would seem. To echo the proclamation of doom: There is no intrinsic, decisive test which could resolve the question, “Are these lines the output of ChatGPT-4?”  Yet, this result is not at all surprising: that the artifact need not ensign the artificer is, of course, what made plagiarism possible before ChatGPT-4, just as it is what makes plagiarism possible through ChatGPT-4.  What, before ChatGPT, made plagiarism most dicey was not source-searching programs like Turnitin, but a history of conversation between student and instructor, and more exactly, a history of shared, conversational inquiry into the text to be essayed.   

An essay which were the continuation into prose of shared conversational inquiry, and the fruit of joint questioning, would be unlikely to blunder into guarded phrasing, shading into falsity (“Emma initially favors the match . . .” when she promotes it strenuously—and ingeniously—right up to Elton’s shocking avowal that she, not Harriet Smith, is his object); it would not weave a garland of vacuously abstract characterizations around that “so good understanding” commended to the author’s particular attention (“alignment of perspectives . . . testament to the power of discourse and respect . . . mechanism for preserving social balance, and a stepping-stone in the evolution (!) of Emma and Mr. Knightley’s relationship . . .”); it would not name-drop Hannah Arendt; would not, in short, consist almost entirely of innocuous padding.

Nevertheless, as my colleague hastened to remind me, one or two more rounds of “chat”—a matter of, at most, two or three additional minutes on line—might well serve to paper over tell-tale traces of a machine-ear for idiom, like “aspiration . . . for status quo”; or of the algorithmic, faux-reading that, above, twists Emma’s visceral disdain for the Eltons’ souless pretensions into an echo of the Borg: “You must be re-integrated”; or even of the persistent implicature that “the Eltons” sum to Harriet and the Vicar.  

If by “essay” is understood, au fond, an artifact, an intellectual property, then undergraduate plagiarism is a kind of conversion or unauthorized, unacknowledged use of another’s property, namely, as a token in exchange for course credit, and (cumulatively) for what is called a credential. Into this context, ChapGPT enters as the unsettling emergence of the custom, counterfeit artifact, pawn of a “plagiarism” sans conversion of anyone’s intellectual property, and thus de facto proof against conventional source-searching, to boot. That context, however familiar, belongs to the illiberal academy, trading in illiberal artifacts, that is, trading in academic commodities, things with conventional uses, instrumental goods. It should be no surprise if students, evading inconvenient convention, seek instrumental goods without the trouble of confecting them. That latter possibility, as Aristotle noted long ago, is the mark of such goods.          

If by “essay” is understood (as Montaigne understood it) a venture, an approach toward the comprehension of the wondrous (and at the same time an offering in aid to all who might share that wonder, since nothing is comprehended unless it is made common in ordinary human speech), then we may mark the essay as one expression of the pre-eminent act of natural intelligence, the which no Language Model, however Large, can enact. That act is the formulation and pursuit of a question, of the unqualified desire for an answer, which lies utterly beyond the realm of machine order. It is the final and serious human business; it is the substance of liberal education, the stuff of human freedom. Wrenching that so many are induced to forego it, to no real advantage, for what compares as a mere “mess of pottage.” 

Love our Articles? Spread the word!

newsletter sign-up

Magnus News

Stay up to date on free courses, webinars, podcasts, and more. 
Bonus! Sign up today and get a FREE At-A- Glance Guide to the Liberal Arts, complete with recommended reading, available for a limited time. 

Subscribe Now and receive a free Guide to the Liberal Arts