The Friction Myth

Part 1: Amplification, Access, and Equity

Image generated by ChatGPT 5.5 High.

Some of the smartest students I ever taught hated writing. The problem was rarely that they had nothing to say. Usually, they had too much: ideas, arguments, jokes, stories, entire worlds in their heads. But the mechanics of getting those ideas onto a page could turn the whole thing into a wall.

Spelling. Syntax. Handwriting. Working memory. The blank-page freeze. The strange humiliation of knowing exactly what you mean and watching the sentence fail to carry it. I started my career as a special education teacher, and I spent years watching people mistake writing fluency for thinking ability.

So, when people talk about AI writing as if the only question is whether it cheapens “the craft,” I hear something else under the argument. I hear a defense of the gate.

Writing as art matters. Sentences matter. Voice matters. I am not arguing otherwise. But there is a difference between protecting craft and treating unaided polish as the only legitimate form of intelligence. That difference matters now, because AI changes the writing surface. Used well, AI can give more thinkers a way through the gate.

I write with AI. Not always, but often, and I have come to enjoy the process. My favorite part is when I come into a planning session with an AI system and brain dump everything I have floating around in my head. I think and write fast and intensely, and having a collaborator that can follow the thread, track the connections, and help me get the strongest version of the idea into print has become part of my writing practice.

I also read more than I write, by a long shot. I read human essays, AI-assisted essays, and pieces generated mostly by AI. Some disclose the process. Most do not. When the content is engaging and it makes me think, I care most about the idea and the work it does.

Apparently, a lot of people feel differently. It feels like every week, another cultural critic or AI alarmist publishes a hand-wringing essay about AI killing human writing. I am an educational leader, so I hear versions of this argument from English teachers quite a bit. And I understand part of the worry. The obvious generated blog posts are bad. The SEO sludge is mind-numbing. The LinkedIn thought leadership was already performative and soulless before AI touched it. AI didn’t make it better.

Those examples prove that weak inputs scale. But bad writing has always spread when it was optimized for attention instead of thought. AI just lowers the cost and turns up the volume.

AI is an amplifier. It amplifies what you bring. If you bring skill, voice, context, and standards, the collaboration can become genuinely useful. If you bring a one-line prompt and no judgment, you get median language at median depth. That is where a lot of the “AI slop” conversation goes wrong. People look at generic output and blame the AI, as if the human brought nothing to the exchange. But the human did bring something: a thin prompt, weak context, low standards, or no willingness to revise.

The amplifier did its job. It made the signal louder. Sometimes the signal was bad.

It does not mean all AI-assisted writing is good. It definitely does not mean writers can stop caring about craft. But it does mean we need to stop pretending that the quality problem lives entirely inside the model. The quality problem often starts before the model ever responds.

And this is where the equity argument matters. Some people are incredible thinkers but not great writers. They can explain a complex idea out loud and then freeze in front of a blank page. They can build the argument in conversation but lose it in the mechanics. They can know what they mean before they know how to make the sentence hold it.

AI can lower that wall. If someone uses AI to express a novel idea, an idea that is actually theirs, one they earned through thinking rather than typing, I am all for it.

Creativity is larger than the sentence that carries it. It lives in the question someone asks, the connection they notice, the risk they take, the story they decide is worth telling, the argument they build, and the impact an idea can have once it finally reaches another person. Sometimes that is one reader. Sometimes it is an entire audience. Either way, the human contribution cannot be reduced to whether the punctuation looked suspicious.

That is the part of the AI writing panic that bothers me most. We have started treating surface features as evidence of humanity. The sentence was too clean. The structure was too orderly. The transitions were too smooth. Someone used too many em dashes. Therefore, the work must be less human, less earned, less real.

That is a strange place to draw the line.

There will be slop. There has always been slop. Bad writing did not begin with AI, and shallow thinking did not need a language model to find an audience. But the existence of bad writing has never been a good reason to stop looking for good ideas.

That is what I do not want us to miss. AI will make some lazy work louder. Of course it will. Every communication technology has done some version of that. But it may also help ideas surface from people who were blocked by the old mechanics of expression: the blank page, the polished sentence, the performance of fluency, the old gate around what “real writing” is supposed to look like.

That is worth taking seriously.

This is a remarkable time to be a person with a thought that used to get stuck on the way out.

So yes, bring standards. Bring taste. Bring skepticism. Bring revision. But also bring curiosity. The slop will be there. Look past it long enough to ask what else is coming through.

In Part 2, I want to talk about the critics themselves: the people who engage with AI shallowly, get shallow results, and then publish those shallow results as evidence. That deserves its own essay.

For now, bring something. The amp is on.

What are you feeding it?