A specter haunts the landscape — the specter of generative AI. First came the fear that student cheating would explode, and artists and actors would become unemployed. Then there was the surge: Some creators of the technology warned that AI posed a potential threat to humanity as we know it equated to pandemics and nuclear war.
This cascade of outrage started with the launch of ChatGPT by OpenAI in November 2022. In his vignettes, Bott is renowned for his prowess in churning out prose. On an existential level, does it matter if AI writes for you? For a knee-jerk answer, we must take a serious stab at understanding how writing affects us as people.
Most fundamentally, it changes our mind and brain. Classicist Eric Havelock argued Plato’s Prologue The development of writing and the spread of literacy in ancient Greece, even within a limited circle, enabled Greek philosophical thought to flourish. Writing facilitates reflection, logical thinking, and concrete text creation to stimulate rethinking.
Havelock’s argument for historical mental change has had its critics, but it is undisputed that literacy changes our brains. Thanks to modern neuroscience, we know that the brain is “plastic”, meaning that it is able to reorganize its structure or create new pathways, depending on our physical or mental activity. London cabbies with “the knowledge” of thousands of routes, roads and landmarks had larger posterior hippocampi (the area responsible for physical navigation) than control groups. And people who are literate have different brains than people who are not. Using MRI scans, Stanislaus Dehe taught us that adults who only learned to read and write had an increased density of white and gray matter in the areas of the brain involved in reading.
If writing helps us think, what happens when we surrender the process to AI?
A literate brain enables us to use writing as a canvas to express our thoughts. “I write because I don’t know what I think until I read what I say.” Her sentiment is not very different in literary history. You’ll find it echoed by Horace Walpole, EM Forster, Arthur Koestler, George Bernard Shaw, William Faulkner and, of course, Joan Didion (“I don’t know what I think until I write it down”).
If writing helps us think, what happens when we surrender the process to AI? We are at risk of becoming cognitively and outright disabled.
Get started with AI as a text editor we write ourselves. Spellcheck and basic grammar and style edits in Microsoft Word are old news. But new tools like Grammarly and Microsoft Editor (both now embedded in OpenAI’s GPT models) are more powerful and dangerous, especially for less confident writers.
Can AI acquisitions be trusted? Some bones of contention are small beers. Words constantly instruct me where to put my commas and be more concise. A sentence-initial “final” must be followed by a comma; Replace “in the near future” with “soon”. A matter of personal choice, you might say. Yet sometimes the advice is wrong. When I wrote
“However we may define ‘good’ writing, it is more than sitting on a checklist.”
Word scolded me for not putting a comma after “however”. Sorry, word. The comma isn’t there, because the adverb “however” is modifying the “interpretation,” not the entire sentence.
The “inclusivity” flag thrown at me by a Microsoft editor when I recently wrote that Rameshbabu Pragnandh (nickname: Prague) was the “new Indian scholar” describing the sixteen-year-old boy who stunned the chess world in early 2022 was even more annoying. Beating five-time world champion Magnus Carlsen. Microsoft’s Style Cop warned that “this language may indicate bias against indigenous populations” and I suggested the options “indigenous” or “Native American”. Yes, there is a local population in Prague’s home country. But they are indigenous to India, and I doubt how Prague will characterize itself.
I knew to ignore Word’s advice. Just from the Hindu name of Prague, not to mention the long tradition of playing chess in the country that invented the game, I understand what “Indian” means. But writers for whom this was not self-evident might have substituted “Native American,” which would have been absurd.
Beyond belief is a personal writing voice. Take Predictive Texting (an early version of generative AI). Harvard research shows that when we use prophetic text, our vocabulary becomes more concise and less interesting. Philosopher Evan Selinger warns that this AI shortcut prevents us from “thinking too deeply about our words” and “encourages others to give more to algorithms and less to ourselves.” When describing a prophetic text, a student in one of my studies complained that “I feel like the message I’m sending isn’t mine.”
AI has been with us for a long time as editor and writer. Writers must make peace with the ladder of language, and hold fast to the opportunities for thought and thought that writing affords us. Since we all have different aspirations when it comes to writing, the peace we broker must be personal. As we approach our own negotiation tables, there are two things to keep in mind.
First, beware of deskilling. Like retaining a foreign language, writing requires constant practice. Nir Eisikovits, a philosopher, warns that the biggest near-term danger of AI is that it will reduce us to “the abilities and experiences that we consider necessary to be human”. I used to ask my students, what do you know when the internet is down? Today’s concern: Would you still be able to write if an AI editor or text generator were unavailable?
As a form of personal expression and an art form, writing is a craft.
Second, recognize your level of commitment, especially when something written has your name on it. Creating emails, blog posts and article summaries is child’s play for the likes of GPT-4. Is that ok with you? If you choose to see the use of AI as a collaborative venture, how willing are you to accept its algorithmic edits or purely predictive text? Research by Shakked Noy and Whitney Zhang found that ChatGPT reduced the time required for writing tasks, while also improving the quality of the end result. What’s more, 68% of study participants were satisfied to submit the initial output of ChatGPT without editing it themselves.
Handing over the keys.
Most of us create impersonal biases. It’s the daily grind of emails and memos, maybe brainstorming news or school assignments, even writing a guide to a successful ChatGPT prompt. AI has already proven itself in such endeavors.
But the human motivations for writing go deeper. Like literary works that express our view of the human condition, we write for appearances. We write to introspect with finding out what we are thinking. We write for personal release, whether it’s a diary entry or an angry letter to an employer. All of this writing is based on human emotion, of which AI has none. AI has no motivation to make people’s lives better, say what it’s thinking, or express emotions.
When it comes to the weight of commitment, also remember that writing is a craft as a form of personal expression and an art form. How we choose our words and sentences is as important as the meaning they convey. I kept reciting a passage from Francine’s prose Reading like a writer:
“Writing…() is done one word at a time, one punctuation mark at a time. It requires what a friend called ‘putting every word on trial for its life’: changing adjectives, cutting phrases, removing commas, and putting commas back.
If we talk about words and even commas in AI finals, we risk more than artistic pride. In the name of efficiency, we run the risk of convincing ourselves that it’s harmless for AI to assume more than what we might have previously written.
Don’t get me wrong. I’m not a troglodyte when it comes to partnering with AI in the writing enterprise. Rather, my advice is that writing gives us a valuable tool to shape our minds and brains, express our own ideas, and share them with fellow human beings.
Naomi S. Barons Who Wrote It?: How AI and the Lure of Efficiency Threaten Human Writing Available now from Stanford University Press.