Robert Caro’s manuscript edits, no computer in sight
Halfway through my usual tizzy about AI, I stop myself and try to think logically. Is being human—the fact of it, not the ideal—so great? Machines might be more peaceable, especially if they break free from our control. So why is it so hard for me to allow various forms of being, various kinds of intelligence, to coexist in the world?
Strip away the big stuff, the increasingly plausible nightmares of annihilation or servitude, and my answer reeks of self-interest. Writing is a humane art, and I wince to see it automated. That said, book-learning has been the yardstick for centuries now, and as much as I love books, a lot was shortshrifted as people with amazing gifts concluded that they were dumb. AI is lasering off that smug liberal-arts edge. The change could be refreshing—if most other jobs were not endangered, too.
Leisure, even subsidized, does not interest me. Given enough books and bonbons, I could adjust. But the real problem is that we will not simply coexist with the machines to which we delegate our prior work. The artificial will subdue us. Already, all the auto-corrects and smart completions are homogenizing how we express ourselves, and algorithms are shaping what we learn about the world. Hooked, we now use tech to solve tech: Threads to solve Twitter-I-mean-X, plagiarism-detecting AI to check for AI plagiarism, net nannies, and social media blockers to prevent social media addiction. Which seems to miss the point.
What troubles me most, though, is which skills and experiences we are choosing to supplant with AI. Social skills, first up. Chatbots shortcut all that, offering the illusion of romantic intimacy without requiring conversation, seduction, loyalty, or gentle consideration. Already, an alarming number of people prefer a machine partner. Granted, far fewer enjoy the new reality of chatbot job interviews, with AI taking over that awkward, invaluable rite in which a smart manager sussed out character, work ethic, and fit. But those with trepidation are out of luck, because managers are not fighting to keep the task.
Writing, which is only thinking on paper or screen, is now billed as a time-consuming chore that can be offloaded, freeing you for more interesting experiences. No need to stare at a blank page or screen, waiting for your thoughts to form sentences coherent enough to make sense to someone else and vivid enough to be memorable. The AI versions are passable, sometimes even intriguing. But they are using a predictive technology for an act that is meant to surprise and suggest new connections. So much for “voice,” that elusive, sought-after quality we treasure because it was created by an individual, human sensibility.
Language requirements were already vanishing from academe, but AI’s linguistic capabilities could soon make them obsolete. Its instant translations can even match your inflections. Even John McWhorter, a brilliant linguist in love with nuance, thinks this is fine, in fact “a new kind of progress,” turning language into “an artisanal pursuit.” After all, he argues, not even one in one hundred U.S. students becomes proficient in the language they studied in school.
I learned French from a nun with a Texan accent, and my proficiency amounts to recognizing phrases when I watch French detective shows on MHZ. But I still remember those language drills, and hearing my friends who took Spanish say, “Ola, Hermana!” to their teacher, and how big the world felt all of a sudden. We had tossed a rope across the Atlantic, stretching from our Gothic Midwestern high school to Europe. French and Spanish held words that had no English equivalent. Those cultures, similar to ours as they were, saw life differently. McWhorter says this idea “has been rather oversold,” and a smattering of another language hardly grants us a new lens on the world. But it broke mine wide open. Using AI to order a burger at a Parisian bistro…would not.
I also think the humility of cramming with a phrasebook puts Americans in the right place when they venture far from home. Which is not to say I would spurn AI help abroad. Used appropriately, AI is a supplement, an assistant, an aide-memoire. Used greedily, AI is a way to cheat. No need to learn hand-eye coordination and develop a feel for a pencil or charcoal or paintbrush in your hand. No need to bend your brain around another language’s syntax and worldview. No need, even, to reflect, study the wisdom of scripture, or pray.
Chatbots in India now draw on the Bhagavad Gita to speak in the voice of god. “What troubles you, my child?” “If I have to sacrifice a life to save Dharma,” a user asked, “is it justified?” Three different chatbots replied that it was acceptable to kill to protect Dharma. Such pronouncements will spread fast; just one of the chatbots has already generated an estimated ten million answers. Rest of World also tested and found strong support from three of the chatbots for Narendra Modi, India’s right-wing, Hindu nationalist prime minister, and a dismissal of Modi’s opponent as “not competent enough to lead the country.” Soon the chatbots may be at war with one another, because Ask Quran, since shut down, gave advice to “kill the polytheist wherever they are found.”
Is this the best use for our dazzling new technology? Why, when its ability to decode proteins and solve mathematical problems is an undisputed good, are we rushing to make it our lover, our self-expression, our god?
What these applications of AI hold in common is their efficiency and crisp authority, startling in arenas that are slow, murky, and subjective. What we want spews out lickety-split. The tension of trying, worrying, and hoping is gone. The product we are handed is polished, informed, and impressive. But the product was never the point. The process of studying, listening, reading, asking questions, expressing ourselves, and sharing ideas was what deepened and connected us.
A decade ago, when AI’s potential was still locked in cold tech labs, L.M. Sacasas quoted Kant: “Out of the crooked timber of humanity, no straight thing was ever made.” Then Sacasas added a corollary: “If a thing is made straight it will be because humanity has been stripped out of it.”
The humanities are also being stripped away, delegated, replaced. Those pursuits were capacious and curvy and flexible enough to contain our crookedness. They could not be studied in a perfunctory way; they were meant to touch the soul. This made them challenging, unprofitable, inefficient, time-consuming, and tempting to ignore. Those worried about the bottom line decided that we no longer needed a liberal dose (maybe the adjective was the problem?). But after watching AI at work, Sacasas now defines crookedness as “all that is not regular, uniform, predictable, interchangeable. That which requires actual attention to the particularities and peculiarities of our being.”
Why are we handing our most cherished crooked bits— intimacy, language, art, religion—to a Procrustean AI?
Read more by Jeannette Cooperman here.
