As my dear friend Laura Kalpakian said recently, "AI is like the snake in the garden...you can all but hear it hissing."

What follows is a series of random observations, at this very specific moment in history, and my attempt to clarify my own thoughts and ethics around this reptilian threat, specifically LLM's—in publishing and in general. (A careful reader will note the use of an em dash in the previous sentence, possibly thinking, hey am I reading an article written by AI? No. This blog post is written by a human. A human with a fondness for em dashes and the rule of three, but an increasing tendency to cringe now when using them, for fear of being accused.)

Author's Guild Guiding Principles

The Author's Guild forum, as you might imagine, is a hotbed of AI discussion. As of May 11, 2026, the AG has released a set of best practices for their members here. I'm certain those guidelines will be shifting as the technology changes, along with society's views on the topic.

My Own Experience

When Chat GPT released its public model, I was super curious. I've always been intrigued by new tech. The environmental impacts weren't part of my awareness, and I did not fully understand the repercussions of how it had all been created. Among other experiments, I fed it samples of polished draft and asked it to create new draft based on my polished style. Several things happened:

• 60% of the resultant text was horrible. (I am sure it would be better now, but I'm not checking.)
• 37% was not horrible, but just not the way I would say a thing. And isn’t voice just a lifetime of struggling—for yourself—toward the way that you, and only you, would say a thing?
• 3% of it was brilliant. So brilliant, in fact, that I felt sadness because I hadn’t thought of it in the first place, and sadness that I could not use it because—to me at least—it glittered like a ruby in the mud of the obvious—clearly not my own. I understand now that these gems had likely been lifted right out of another author’s book, and I was right to be suspicious.
• Lastly, I had to examine what the experience of writing is all about, and—for me, at least—it isn't about producing text.

When I was in art school, a visiting ceramics professor was rumored to have asked students to bring their finest thrown pots for critique. When everyone had assembled, he pulled out a stick and smashed everything, accompanied by tears and wailing. He then pointed to the shards and said, "These are only artifacts, not the art itself. The art is in you. The art is in the doing."

That story has stuck with me for decades. Art is not a product. Art is in the doing—in the transformation of the individual wrestling with the process. The alchemist is the one changed.

An hour of generating AI text can produce many thousands of words. By contrast, I might spend an hour reaching for the exact right phrases to catch an ephemeral moment that only I—as a unique individual at a singular cross-section of experience—could ever convey, and I'd be thrilled to "produce" 500 words. But I will be forever changed and enlivened by the process. The way I go on to notice my experience deepens—becomes richer for the noticing—thus changing what I experience and choose to convey in future. And so on, in an ever-spiraling feedback loop of transformation. The very best artistic expressions create a similar experience for the one who reads/listens/watches. AI will never do that.

Beyond that, AI just can't do the thing we need art to do. It is a homogenizing tool, meant to find the most likely next word. We look to art to surprise us, to do the thing unexpected, to push the boundaries. I named this blog post Parasitic Mushrooms, Snakes, and the Chateau D'if out of a unique intersection of memories and associations that made sense to only me, then tried (perhaps failed, but at least tried) to convince you of their relatedness. I ran this post through Claude and asked it to suggest a title: AI in Publishing: One Author's Principles, Conspiracy Theories, and Hard Lessons.

Accurate, but...oh my goodness, no.

To be transparent, I use AI in my day job. Finding the missing semicolon in 1200 lines of code is a mental task I am relieved to turn over to the machines. I do not want my code to be creative. I want it to be deadly average. But, having wrestled with its impact, I limit its use where I can. I want my mental and physical ecology preserved. I don't want it in my literary life, beyond the grammar checks in AutoCrit. What's more, if I were trying to get rich by using AI, I certainly wouldn't use it for writing! I'd spend my time creating agents to predict the NFL draft or invest in the stock market—MUCH better ROI.

The Nobel Prize Drama

The recent experience of Nobel Prize Winner Olga Tokarczuk suggests the current guidelines published by the Author's Guild are not in line with public sentiment. Or, at least, the public-facing sentiments of the literary community.

In short, Tokarczuk gave a talk in which she admitted to using AI to learn what dances people might have been enjoying at a particular time in history. She also mentioned using AI to refine her prose. The literary world went wild: articles were dashed off in fiery indignation, writers took to social media to excommunicate her, pitchforks were raised and torches lit. She went on to clarify that she did not use text generated by AI in her finished manuscript, but her village was already blazing. The Author's Guild is American, and Tokarczuk is Polish, but—using the AG as a stand in for other, similar organizations in Europe—her actions do not seem to breach the conditions set forth in principles #2 and #11.

Curious about the situation, I turned to an article in LitHub, in which Tokarczuk's original speech was translated by the LitHub author using Google Translate. She pointed out the irony of criticizing Tokarczuk for using AI while writing an article aided by AI translation. In my view, self-admitting the hypocrisy does not expunge it, but it does point out the difficulty: AI, like a toxic mushroom, has invaded everything that we do on the daily, to the point that our dependence on it is becoming an invisible, if untenable, symbiosis. Twenty years ago, the author might have found a Polish translator. But with the prevalence of AI translation tools, who can justify the expense, publishing being what it is today? (To say nothing about what AI is doing to translation as a profession.)

Still curious, I clicked on another news article covering the drama. This one was also highly critical of Tokarczuk's choices, but turned out to be AI slop, without even a brush of humanizing or an author credit at the end—an AI-generated article excoriating authors for using AI, which got me thinking about...

Conspiracy Theories

Writers are the very people whose treasury has been raided to create LLM's. Every sentence created by a human mind is like a flake of gold in the dragon's hoard. Yet writers are currently the only population most likely to be tarred and feathered for use of AI. (To be clear, I am not advocating that writers begin vomiting generative AI into their manuscripts, I am just pointing out a curious dichotomy.) Why are writers, the population whose work has been pillaged to create this golem, the only ones who will be pilloried for using it? It seems wrong for every other industry under the sun, from medical research to auto racing, to benefit, while the very people who were robbed to create it, are not allowed to benefit. Where is the justice in that? It's like being criticized for buying back your own stolen family heirlooms—now scratched and battered—at a pawn shop, because the authorities won't recognize that a grievous theft has occurred.

But what if...what if...(donning tinfoil hat) the tech companies themselves are adding fuel to the fire. Encouraging the literary community's self-policing in the world of AI to keep them churning out sentences?

When writers attack other writers for using AI, tech companies win:

• Attention stays on "bad individual users" instead of extractive corporate practices. It's giving shut up and use your paper straw, Bobo, while we dump sewage into the Everglades.
• Writers remain divided and powerless but (critically) STILL PRODUCING HUMAN-AUTHORED SENTENCES to be devoured by a monstrous, starving ouroboros.
• The focus stays on who is using AI instead of
• What compensation authors deserve, how these technologies can be regulated, how to organize into collective action.

The Chateau D'if

AI detectors (created by AI) are increasingly finding AI tells in recently published works. Accusations fly. How can people insisting on purity tests against AI keep a straight face while using those same tools to check? Muddying the issue, AI prompts are easily written to avoid AI tells detected by the detectors, and so on. Reminds me of the Star-Bellied Sneetches.

I do wonder, too, as we all engage in more and more content created by AI, do we, as writers, tend to mimic those tells? I was already an overly-caffeinated user of the em dash, and the rule of threes is what feels natural to me, as a writer and a reader both. Will AI's stylistic quirks, like the Green Grocer's Apostrophe, begin slipping naturally into our writing? I cannot tell you the number of times I've had to delete an apostrophe from a word meant to be a plural—a mistake I would NEVER have made before decades of internet use rotted my punctuation.

I also worry about the freedom with which writers are accusing one another of using AI. One indignant judgement by a high-profile author on social media can mobilize the swarms, who are always eager to join a righteous take down. The recent controversies surrounding the Commonwealth Prize and Shy Girl highlight how quickly a writer can be cut down to the anklebones or tossed into the literary equivalent of Chateau D'if with poor Edmond Dantès.

Again, I am not defending AI usage or claiming any stance on the above two situations, only pointing out how frightening (and disheartening) it is to see writers willing to destroy each other's careers without proof or nuance. The fact that I keep needing to inject disclaimers into this post is evidence of how truly spooky this has become. I'm fearful that pointing toward any ironies, hypocrisy, or shades of gray will put a target on my head.

The End Game?

The tree of evolution is filled with dead end branches. Among these are the unsuccessful parasites that kill their hosts too quickly. I worry that AI, like the toxic mushroom I mentioned before, will grow into every aspect of our human lives, killing all that it touches, and then die out itself, leaving a wasteland.

AI tech companies are already drinking salt water to slake their thirst:

1. They devalue human creative work (economically and culturally.)
2. This kills human creativity (why write if AI does it cheaper?)
3. Less human creativity equals less training data
4. Future models degrade from being trained on their own outputs.
5. The outputs become less useful
6. And so on, down to the bottom.

In short, this parasite is destroying the very thing it needs:

Us.

*****************************

ADDENDUM: 06/16/2026

We live on a beautiful little creek surrounded by woods. Have recently learned that a massive power line will be built through the creek, land seized by eminent domain to feed the beast. What's more, our little rural community is being scouted by builders of the big data centers. If any of us think we won't be impacted by this monstrosity, we are wrong.

I am heartsick.