The Rite of Authorship

On AI and the right to embodied voice

“Nothing without courage.”

This week, waiting to board a plane home from a work trip, I read an article in The New Yorker, “Can A.I. Produce Writing That We Actually Want to Read?” The author, Jay Caspian Kang, created an experiment: train AI on the voice of several famous authors; ask it to produce texts in the style of those authors; then ask humans to parse out the originals from the counterfeits. At first, AI fakes were easy to spot. AI-generated characters “apart from fidgeting… mostly did nothing.” When Kang pushed Claude to produce more decisive characters, Claude took it too far. Every character sprang into action, jumping on horses, delivering urgent messages, or running. As Kang’s experiment continued, the fakes improved, ultimately with as little as 17% of testers discerning the difference between real Bram Stoker and AI Bram Stoker.

Kang concluded by drawing a parallel between the writer’s relationship with technology and that of grandmaster chess players. Computers have reliably beaten humans at chess for decades; yet, humans still pursue the game. Humans play chess because they value the cognitive process of the game, even if computers can smoke us. Kang argues that writing is the same. People will still choose to write as a mechanism for synthesizing thought and communicating with others.

The article’s algorithmic arrival at the top of my newsfeed was timely. Like so many, I’m fatigued by receiving and parsing AI-written memos—rife with overly ambitious word choices, em dashes, uncharacteristic cheer, and the most annoying telltale marker of all: “it’s not X; it’s Y.” My brain is tired. Increasingly, I find myself drawn toward two co-existing poles: transparency in the use of AI, and the raw beauty of messy, human authorship. Kang’s article touched on a topic I’d been circling for weeks. I wanted to learn more, but wasn’t sure where to start.

So, I turned to Claude.

I asked for additional article recommendations. Claude replied with a few strong contenders, including academic papers and Substack authors I had never heard of. I spent the better part of my plane ride reading additional perspectives.

When I came home, I told my husband about the articles. He responded by sharing a few more.

Among many authors, a central theme persisted: the human process of writing is best left unassisted by AI. João Batalheiro Ferreira expressed skepticism towards leveraging AI in the brainstorming or revision process. Allan Reeder referenced his experience as a writing coach who guides college applicants in essay authorship, arguing for the precious value of slowing down to wait for original, true insight and “what we call voice.” Eve Fairbanks lauded the inefficient hang-ups and backtracking of the writing process as “a feature of human thought process, not a bug.”

I’ve known many writers obsessed with the process. My ex-husband—a fantastic writer, gifted with skill, smarts, and three college degrees—used to insist he could only write in a room with hardwood floors. Everyone has their thing. In many writers’ eyes, the blank page is a solemn rite of passage. A mountain every author must climb.

But here’s the thing: the page was never sacred.

Of all the articles I read, the most fascinating was one Claude suggested, an academic paper by Jon Ippolito. Ippolito argues that authentic voice—that elusive, uniquely human, deep soul truth—has been augmented for thousands of years. Those who argue that AI is offloading a cognitive process aren’t wrong; but, they are missing the many historic inventions that progressively altered the cognitive process of expression. The alphabet shifted traditions of carrying stories through song, oration, and memory. The Gutenberg press shifted the perception of calligraphy as essential to expressing personality. Word processors shifted the editorial process, offering a digital means of revising the written word. “Perhaps the loss of authorial personality predates AI-generated prose; perhaps the slippage began the moment the alphabet separated the speaker from what is spoken, and the gap has been widening with every introduction of a new writing medium.”

Ippolito acknowledged those who valorize the writing process as a mode of self-discovery and cognition. But, he argued that we might be on the precipice of a new evolution of human expression, a new split: between discursive, oral, or embodied personal communication and compact, machine-readable, impersonal communication.

I see traces of Ippolito’s argument in my own process, which I began well before reading these articles. I’ll cut to the chase: my process includes AI. It does offload certain forms of cognition. But, perhaps not where you’d expect it.

Here’s how it works:

• Step 1: Wind aimlessly around ideas spinning in my brain. Question the worth of said ideas.

• Step 2: Sit alone in a quiet room or parking lot. Record voice memos, as if talking to my closest friend. Gesticulate. Sigh. Pause. Wander. Stop the recording. Get back to life.

• Step 3: Upload voice memos to Claude. Ask Claude to interview me. Answer Claude’s questions in writing or voice.

• Step 4: Ask Claude to stop blowing smoke up my ass when it tells me I’ve hit on an insight “nobody but you could have thought of.”

• Step 5: Read Claude’s observations of patterns and connections. Correct misunderstandings. Request a suggested outline. Discard half of it.

• Step 6: Draft, in my own words, outside of Claude.

• Step 7: Upload a draft to Claude. Remind it (again) to stop being a sycophant. Ask for input on what’s structurally sound or weak.

• Step 8. Go back to my draft. Revise. Revise again. Revise again. Revise again.

• Step 9: Feed my final draft to Claude. Request input on fact checks and grammatical errors.

• Step 10: Fix. Publish.

My process doesn’t begin with a blank page. It doesn’t include a writing coach. It doesn’t include an editor or a publishing house. It doesn’t require hardwood floors.

But, it does take time. Time borrowed in between the competing demands of everyday life. Hours spent typing away in a bathrobe on a Sunday morning. Stepping away, coming back. Reconsidering. Restructuring. Ruminating. Resolving. It’s not a diamond mined from the earth. It’s grown in a lab of voice, written word, an overly eager, unpaid AI writing coach, and more espresso shots than I care to enumerate. But it’s a diamond nonetheless.

Some might say I’m skipping the necessary friction. I disagree. The friction remains. It just lives elsewhere. It lives in the part of me that questioned starting a Substack. It lives in the moments when I criticize the value and worth of my ideas. It lives in the hours spent writing and rewriting. It lives in the pernicious voice under the surface when I hit publish, asking whether it’s “good enough.”

My process has removed the fear of trying.

Embodied in voice, hands, words on page, my process is my own.

My process isn’t perfect. But it’s no less true.

In a world oversaturated with cheery, tidy, AI-written prose, I want to hear the words and voices of other humans. I want to hear the messy thoughts. The half-baked ideas. The unresolved questions. I want to hear those thoughts regardless of where they began—a terrifyingly blank page, a static-laced microphone, a tentative dance step, or a ten-word AI outline. No matter their genesis, these expressions are brave. It takes courage to share the ideas living within you.

My Substack has given me a place to write, create, explore, reveal—with no goal attached. It gives me a place to let my thoughts breathe. Sometimes, I find out that it resonates with others. I get an email, a text, or a Teams chat from someone who read a post and felt compelled to reach out. They say they felt seen. These messages are authentic. They forge unexpected bonds and connection points. Often, these are some of the most deeply human moments in my week.

They say to trust the process. I will trust mine.