Rereading ChatGPT: Finding Hope in Literary History

Rereading ChatGPT: Finding Hope in Literary History

Jennifer Lieberman | Tuesday, Apr 18, 2023 |  AI MLAutomation

The release of artificially intelligent tools for public use has spurred a heated debate in the past few years. While AI art was a lark for some and an opportunity to experience gender euphoria for others, to artists it represented stolen and uncompensated labor.

In the context of academic writing instruction, the controversy surrounding Large Language Models (LLMs), such as ChatGPT and Bard, has less to do with direct theft; there are no blurred author’s signatures hidden amid the narrative collages these tools provide. Nonetheless, some have suggested that these tools will undermine writing instruction or the study of the humanities more broadly. This existential fear has inspired troves of teachers to pivot from a pedagogy of care to policing, threatening students with AI-detection software and harsh penalties. Meanwhile, others have replied to this fear with suggestions for embracing LLMs. Like many teachers, I’ve read these discussions with interest, wondering how these tools might call for new approaches to the craft of assignment-writing.

As I read, I noticed that the question of authenticity underwrites every piece, although it is often left unstated. The urgent question seems to be: Now that essays can be artificially generated, how can we determine or foster our students’ “real” skills? This public debate filled me with a sense of déjà vu I could not name for quite a while.

Because the name I was searching for was Henry James.

Stuffy almost to the point of unteachableness, James isn’t the type of historical figure who readily comes to mind when reflecting on cutting-edge technology. (Walter Benjamin’s essay on “The Work of Art in the Age of Mechanical Reproduction” has been a more obvious reference, for instance.) Yet James was obsessed with authenticity. He helped to craft an entire artistic movement of literary realism because of this obsession. And I believe he and his peers might help us ask new and generative questions in our current moment, and, in turn, that AI can help us think about his legacy differently.

James and the literary realists rose to cultural prominence during a moment of social upheaval and technological change in the U.S., the likes of which was recently dramatized in Julian Fellowes’ HBO series, The Gilded Age. Even readers unfamiliar with media from or about this era have likely heard the term nouveau riche and might recall how that designation raised questions about prevailing hierarchies of human worth. Who would determine a pecking order among elite white Americans, as wealth diverged from the genteel notion of “good breeding”? These were the questions that plagued James and his contemporaries.

Today, we can see how the very notion of gentility positions eugenics within fuzzy discourses about authenticity — not unlike how today’s LLMs recapitulate biases while presenting to the lay public as new and neutral tools. Many scholars — including myself, but much more notably Safiya Noble and Emily M. Bender and Timnit Gebru et al. — have written elsewhere about how technologies recapitulate and repackage existing biases while seeming to promise social change; that is not the subject of this essay. Here, I wish to explore how ideas about identity became tied up with questions about technology and “realness” in historical examples that resonate today. Edith Wharton — a writer with both wealth and an impeccable family line — wondered if the truly genteel might be better discerned by candlelight than under an electric bulb. James questioned if electro-plating made it impossible to tell whether a tea service was real (inherited and expensive) silver or a cheaper yet indistinguishable set. Sometimes, the connection to technology was filtered through capitalism: the fortunes decided by the booms and busts of railroad speculation, for example.

The short story that addresses the theme of threatened authenticity most directly is James’ “The Real Thing” (1893). In this piece, a genteel husband and wife, transparently named Mr. and Mrs. Monarch, arrive at an artist’s home. He first believes that they want to commission a portrait, but he soon learns that they hoped to earn money by modeling for him. This miscommunication is the first of many.

The difficulty of defining the titular “real thing” is ultimately the point of the story. The Monarchs and the artist can agree that “real” ladies and gentlemen have “it,” but they cannot express what “it” is or what it means in their present moment. We see both this theme and its elusiveness when Mr. Monarch tries to convince the artist to take them on as models by stammering:

‘Wouldn’t it be rather a pull sometimes to have—a—to have—?’ He hung fire; he wanted me to help him by phrasing what he meant. But I couldn’t—I didn’t know. So he brought it out, awkwardly: ‘The real thing; a gentleman, you know, or a lady.’

The artist agrees at first, but their shared hypothesis fails. After trying to draw both husband and wife, the artist discovers that the Monarchs are too authentic to be anything but themselves. His servants prove to be better models. After he breaks the news that he cannot employ them, the artist tells the reader, “They had bowed their heads in bewilderment to the perverse and cruel law in virtue of which the real thing could be so much less precious than the unreal.” The story ends with a reversal: the Monarchs and servants switch roles, with the poor-but-genteel characters accepting positions as servants, while the former servants take up the mantle of models.

James’ story raises questions about how cultural notions of authenticity are complicated by a market economy. (Fun fact: the demystification of gentility is one of the reasons capitalism and democracy seemed related in an early American context, although that relationship seems much muddier now.) At the beginning of this story, the Monarchs seem fully invested in the “real” value of their gentility. Yet their need to participate in a market reveals their investment in gentility to be at best, outdated, and, at worst, an illusion. Authenticity is also an existential question for the artist; he would rather create “real” art than illustrations for magazines. New technologies and economic realities have challenged the very notion of reality for these characters. Couldn’t the same be said for many of us, today?

This story may seem far removed from our current moment — except, perhaps, for the schadenfreude of watching the fall of former elites and implicit critiques of capitalism, which can still resonate in our moment of starkly increasing wealth disparities. Still, I believe we can learn a lot from pieces like this one. I find it grounding, for instance, to remember that our generation isn’t the first one to encounter this kind of moral-technological panic. Recognizing this fact can help us to reflect upon our own often unquestioned notions of authenticity and value.

James was immersed in and responsive to a changing moment in economic and labor history, just as we are today. And he adapted his voice accordingly. “The Real Thing” is full of miscommunications, evident in the quotes I excerpted above. He responded to social and technological change by making his prose more difficult, drawing attention to a fact communication theorists would articulate much later — that humans are always miscommunicating, that we can almost never convey precisely what we mean. (There might be an irony here, that while LLMs attempt to create clarity, there will always be an important element of human communication that relishes obfuscation.)

In the past, I’ve read James’ tangled prose as a kind of resistance to technological change, much like the resistance we see when teachers strive to create assignments that can befuddle AI. I had imagined that James’s prose subverted the oversimplified concision of a telegraphic message with artful meandering. But the theme of miscommunication strikes me differently in our present moment, when we have so many barriers that inhibit communication outside of our media silos.

Rereading James today, I suggest that “The Real Thing” raises larger questions about how we understand ourselves and relate to one another. The Monarchs had to radically change the way they related to themselves and the artist; they had to give up their notion of superiority and serve him. And while this change didn’t thrill them, it wasn’t entirely tragic either. Finding new ways of being is hard, but it also signals growth.

Artificial intelligence does pose a real threat to democracy and to professions, including and beyond writing instruction. As a recent petition to pause AI research published by the Future of Life Institute has claimed: “AI systems with human-competitive intelligence can pose profound risks to society and humanity, as shown by extensive research and acknowledged by top AI labs.” But, like Roger McNamee, I’m uncertain if a six-month moratorium could subvert these problems. Instead, as I re-read James, I wonder if the threat AI poses follows largely from the fact that both jobs and democracy were already under threat — partially because of intentionally curated miscommunication. (Would deepfakes be so frightening if people weren’t profiteering from political divisiveness and conspiracy-making? Would they pose the same threat if we were better able to have difficult conversations with neighbors and deliberate policies for the greater good together?)

All of these questions call to mind Gabrielle Lyon’s recent statement: “The public humanities are not in crisis; they are what we need to respond to crisis.”

And that brings me back to James. Like the Monarchs, the artists, and the servants, we are living in a moment when we might need to question intrinsic biases and re-learn how to relate to each other in order to survive. James’ “The Real Thing” illustrates how a fascination with authenticity can distract us from the other threats to our growth. If the Monarchs clung fast to their idea of realness, they would have starved to death. They had to realize that their obsession with authenticity (and, implicitly, hierarchy) inhibited them from change. By realizing that real-ness was changeable or farcical, they were able to re-orient to each other, even creating a small community of mutual aid, where each participant (artist, Monarchs, servants) contributes what little they can to help the other. This model has a message we can learn from today.

Regardless of whether AIs can write papers, writing and humanities classes still inherently matter, because they help learners communicate how they understand themselves and the world. These courses foster reflective thinking students will need when they adapt to inevitable change — whether that change is a loss in status wrought, in part, by the rise of AI or more common (but no-less-cataclysmic) changes in identity or health that are an inevitable part of life. But, as James implies, our fear of artificiality might distract us from the more meaningful work of exploring how hard it is to understand different perspectives and communicate our own.

Rereading James in our current moment, I propose that the “real” threat isn’t a new LLM students might use to cheat, or even the larger social harms that might be exacerbated by AI, but the pre-existing social conditions that invested heavily in the creation of those tools while disinvesting in students — and in education as an institution — for generations. We have de-funded the humanities for decades, imagining that we could fund STEM without any regard for the humanistic or sociological stakes of science and technology. As long as we remain distracted by debates about authenticity, we can miss the more urgent issue: we deeply need the humanities and social sciences to help us build more ethical technologies and to help us cope with socio-economic conditions that new systems, including AI, exacerbate. This can be a comforting realization, because inventions cannot be un-invented, but policies can be re-written. Re-investing in education and the interdisciplinary humanities might be the first step we need to reducing harm for the next generation.

Jennifer L. Lieberman is an associate professor of English at the University of North Florida (UNF) and the Director of the Office of Interdisciplinary Programs, overseeing Africana Studies, Interdisciplinary Studies, and International Studies. She is the author of Power Lines: Electricity in American Life and Letters, 1882-1952, which is available from Amazon.

Photo via Flickr Commons from the U.S. National Archives.

About This Post

A Humanities professor offers some unexpected new perspectives on ChatGPT inspired by the American Realism literary tradition.

Written by:

Share this post:

Recent Rotations butterfly

View all

To LLM or Not to LLM (Part 2): Starting Simple

Sick of hearing about hyped up AI solutions that sound like hot air? 🧐 Let’s use boring old ML to detect hype in AI marketing text and see why starting with a simple ML approach is still your best bet 90% of the time.

Building an AI Text Detector - Lessons Learned

The LLMs boom has made differentiating text written by a person vs. generated by AI a highly desired technology. In this post, I’ll attempt to build an AI text detector from scratch!

May 15, 2024

To LLM or Not to LLM: Tips for Responsible Innovation

We’re seeing a proliferation of Large Language Models (LLMs) as companies seek to replicate OpenAI’s success. In this post, two AI engineers respond to LLM FAQs and offer tips for responsible innovation.

Enter Your Email To Subscribe