Skip to Content, Navigation, or Footer.
The Daily Cardinal Est. 1892
Monday, March 18, 2024
ChatGTP.jpg

The stories machines can’t tell

The concept of artificial intelligence first punctured my juvenile mind on a brisk fall morning five years ago. At the time I was an arrogant ninth grader, enveloped with a façade of self-importance and completely disinterested in the slow crawl of societal advancement.

Community is my suburban high school’s sales pitch, and it’s expressed through the ritual of morning assembly. Every day at 8:15 a.m., our 600-person student body filed into the auditorium — usually to entertain mundane announcements and chidings over our inevitable shortcomings. But every so often, we were treated to the spectacle of an outside speaker sharing “real-world knowledge.”

I especially hated these mornings. Between the hours of eight and nine, all I could think about were the fluffy pillows and warm duvet I illogically abandoned an hour earlier. I groggily lacked the desire to explore any existential questions deeper than what the cafeteria was serving for lunch sixth period.

One colorful November morning in 2017, I deserted my thick pyramid of blankets and collapsed into my assigned seat — hoping to capture the final fumes of sleep before they evaporated forever. I was completely unaware that half of Silicon Valley would have mortgaged a limb to trade places with my slumped corpse.

Our guest that day was Sam Altman, John Burroughs School class of 2003 and President of Y-Combinator, a company I did not know or care about. But during his presentation, he pivoted to discussing a side project — his new start-up, which he called OpenAI.

Five years after I almost fell asleep listening to his presentation, Altman and his little company would release the most powerful computerized assistant ever — ChatGPT.

That vignette is how I, a human being, decided to open this column. From the time we were restless elementary students, our English teachers taught us to “show, don’t tell” because stories engage the audience’s mind. Starting an essay with an anecdote is an effective way to hook the reader, which is why so many writers rely on the trope.

There are many ways to write a lede, of course, because elegance and effectiveness are subjective, highly situational choices. Another route I could have taken was this:

As we move into the future, it is clear that artificial intelligence will play an increasingly central role in our lives. From self-driving cars to virtual assistants, AI is already beginning to revolutionize the way we live and work. But what does the future hold for this rapidly-evolving technology?

Some predict that AI will eventually surpass human intelligence, leading to unprecedented advances in science and medicine. Others worry about the potential negative consequences of such a development, including job displacement and the loss of human autonomy.

As we navigate this brave new world, it will be important to carefully consider the implications of our choices and to ensure that the technology is used ethically and responsibly.

Bored? Yeah, me too. That’s how ChatGPT would have started the article. More specifically, that’s how ChatGPT thought Joan Didion, my favorite writer, would have started the article.

Enjoy what you're reading? Get content from The Daily Cardinal delivered to your inbox

A language master, Didion possessed the command of a Cy Young winner. Her pitches were powerful verbs and descriptive metaphors, which she used like a devastating slider on an 0-2 count.

She once likened the 1960s American to San Francisco’s Grace Cathedral, declaring that “In a city dedicated to the illusion that all human endeavor tends mystically west towards the Pacific, Grace Cathedral faces resolutely east.” When covering the 1988 Presidential Conventions, she mused that there was a reason “the people with whom I had preferred to spend time in high school had, on the whole, hung out in gas stations.”

If I attempted to emulate the woman who titled her anthology “We Tell Ourselves Stories in Order to Live” with writing as vague as what ChatGPT spat out, I don’t even think my mother would have pinned it to the refrigerator.

This is the essay of someone who just found out it's due tomorrow. We’ve all been there, losing it in the library — lacking any substance and needing a whole lot of makeup. We resort to stringing synonyms together (“ethically and responsibly”) and using overworked clichés (“this brave new world”) that once meant something before they solidified into an uninspiring sequence of words.  

Didion often described writing as a “hostile act.” Her job, as she saw it, was to force the audience to understand her perspective through rhythm and unique structure. Channeling this emotion is even more important today. With Netflix and YouTube and so many other competing forms of entertainment, a reader will simply find better things to do than ponder dull redundancies like “already beginning to revolutionize…”

Altman concedes that his technology is still “incredibly limited.” But, artificial intelligence will improve and has improved even since I started writing this piece. Where it is now is not truly important — what matters is the ceiling.

Five years ago, when ChatGPT was no more than an idling concept in a spiral notebook, potential is what Altman focused his presentation on. “This is going to change everything about everything,” he declared at the time.

ChatGPT is a language learning model — it analyzes a slew of written text from every corner of the internet and picks words based on prior human choice. It studies what has been done before and mimics it, not truly creating anything new.

On a surface level, human creativity is not much different. Just as Didion influenced me and my writing, she found inspiration in Ernest Hemingway and Henry James. We do not conjure ideas from our mind’s abyss; we build bridges on pre-made foundations.

Take William Shakespeare, the grandfather of English vernacular, as an example. While he did not invent the rhythmic concept of iambic pentameter, he is remembered as the meter’s master.

He saved it for his most powerful lines, and littered scenes of intense emotion with the style. 

“But soft, what light through yonder window breaks,” Romeo breathlessly utters upon seeing his love basked in the moonlight in “Romeo and Juliet.” “It is the East, and Juliet is the Sun.”

Hamlet, distraught over his father’s recent death and his mother’s hasty remarriage to his uncle, cries out in despair, “Oh that this too, too solid flesh would melt / Thaw and resolve into a dew!”

These soliloquies land not solely because of the iambic pentameter, but because of how Shakespeare used it. The meter is characterized by an alternating rhythm of stress. Read the passage from “Romeo and Juliet” out loud and notice how you naturally emphasize every other syllable. “But soft, what light through yon-der win-dow breaks.”

Shakespeare realized the pattern (duh-DUM, duh-DUM, duh-DUM, duh-DUM, duh-DUM) mirrors the beating heart. Using it in such a romantic scene makes the line especially beautiful — the lovesick Romeo is not just speaking from the heart; he’s speaking with it too.

ChatGPT cannot do this. It can explain what iambic pentameter is and it can describe the pattern of a heartbeat, but it could never build the bridge between the two. When humans are described as emotional creatures, as opposed to rational ones, it is not meant as a compliment. But it should be. Emotions unlock the ability to see the world as a vibrant place. A computer, which operates on infinite strings of zeros and ones, misses the complexity of everything two and up.

Altman understands this too. Back in 2017, he declared building OpenAI “in a way that is good for humans” his primary focus. AI is not here to replace humanity — it is a tool to assist us.

Yes, perhaps one day artificial intelligence will write a perfect, grammatically correct essay. And it probably will be able to follow basic art technique and create a stunning digital image — if it can’t already. But we do not measure excellence in these disciplines by how well artists follow existing conventions, we value the ability to break them in new and innovative ways. Nuance is beautiful, and beauty is subjective. There is no binary in art.

Shakespeare is not taught to every high school student because he used prepositional phrases correctly. Understanding the rules makes you talented and opens the doors to future jobs. Understanding the rules and doing something else, as Shakespeare and Didion often did, gets your name etched in textbooks because it’s impossibly human.

“We tell ourselves stories in order to live,” Joan Didion wrote in “The White Album.” “We look for the sermon in the suicide, for the social or moral lesson in the murder of five. We interpret what we see, select the most workable of the multiple choices.”

AI can help humans push the ceiling higher, but like the printing press, steam engine and Google, it cannot do it alone. Technology helps us develop stories, but we must create them. There’s no engineer who can build bridges of meaning between disparate shores like a human.

Graham Brown is a Sophomore studying Political Science at the University of Wisconsin-Madison. Do you agree that artificial intelligence will eventually hit a ceiling? Send all comments to opinion@dailycardinal.com.

Support your local paper
Donate Today
The Daily Cardinal has been covering the University and Madison community since 1892. Please consider giving today.
Comments


Powered by SNworks Solutions by The State News
All Content © 2024 The Daily Cardinal