Hello my friends and occasional relative. I submit a new rant, which, upon reflection, I possibly should have written in meter. But I didn’t, because how easy would it be to train an AI tool to do that?

Ahem. Tap tap tap.

In my last outing I went straight to the end of the road, the world where AI write our entertainments, and we believe them, and they are good. 

There are a few steps to get there and ChatGPT3 is about three steps ahead of where those of us not paying attention would like us to be.

Analogue is dead. Digital is king. Not if you are an audiophile, but how many of us are true audiophiles? Not very many. The rest of us love digital, and live in the digital aural world. Whether we know it or not.

We are, I am afraid, in the last days of analogue writing. Writing by humans for humans, without tools that all too eerily mimic human capabilities, whilst still being mindless themselves.

And no, a spell checker is not analogous to an AI, nor is a grammar checker, until it begins to substitute sentences for you. Something that points out a possible mistake is not an AI.

ChatGPT (see my links in the previous blog, it’s all there) lets writers work out their plots, writes descriptions of characters, or places, can do passages of dialogue, all from a series of prompts.

The series will get shorter, the AI better attuned to what the right answers are. What we want. Not because it knows, but because of what we, on aggregate, respond to positively.

That is what it takes in as new data. Whichever version we are talking about. And while at present we think we are the masters of the tool, do we ever think about what the tool might do to us?

For now writers who use AI think it is just to do the bits they don’t like, the boring descriptions, the clunky dialogue they haven’t mastered.

It just fills in the bits the “writer” hasn’t the patience, or, in my mind the craft to do for themselves. And it can do it in multiple styles, you know, so our “writer” can find their voice.

My hatred for this phenomenon knows no depths. I have a facility for language, a dexterity with words that is all my own, and now it can be copied, aped, by a tool. Used by children.

Everything in my life that led me to be able to express my thoughts as I do, in the way that I do, can be if not outright copied, at least mimicked satisfactorily to an ignorant audience.

And I believe these tools will make us and our audience more ignorant. The lowest common denominator will rule, a button mashed until it doesn’t work, then the next button will be hit.

Writers who now believe speed is a virtue, who crap out novels at a mile a minute will benefit short term from these tools: they will write, and make, more.

But if the AI is organizing their plots, if the AI is describing their characters, and the environment through which they move, helping with dialogue, what is left?

The writers claim they know the inner lives of their characters, they have mastery of the trajectory of the story, that well, 51% OF THE WRITING WILL BE THEIRS.

How quaint the lies we tell ourselves are. How ready we are to sell ourselves, piece by piece (or bit by bit to be more accurate), just for the illusion of success. For silver and gold.

In the articles I referenced in the last blog we already had a writer admit her beta readers liked the AI writing (in places) more than hers. Do you keep it or bin it after that? Change something and claim it as yours? What a horrible conundrum.

When that AI derived writing makes more money, how long before some writers of lower moral character decide to take credit for it? To go beyond 51%? They are still directing the tool to do what they say, right?

So “writers” become something less, not the writers of the words, but the giver of prompts to a mindless machine that can write with far more felicity than most writers could ever dream of.

So what, or who, drives the public conversations of the future? Endless plausible AI pieces to convince attention starved screen junkies to give just one more second of eyeball time to that particular window?

And this is my fear: that the children born and raised by screens will have no idea of analogue writing, will think writing by tool is natural, right.

If we hand over writing the boring bits, the descriptions, the dialogue, all the filler between the exciting stuff we always wanted to get to as writers, then have we really done our job?

And what if our “fans” like the filler more than our tiny proportion of exciting stuff (I mean deathless prose)? We will end up with the human parts being the weakest. Until the AI touches it up to match the rest. Sweet.

But you, my friend, will still be a writer. Like any other doofus who woke up with an idea and farted it out to Alexa and asked her to arrange it in the style of a space opera by Asimov. Or Dilbert. Or Camus. Or Steele. Or Patterson.

This is the problem. Every idiot will reckon themselves a “genius” writer because the AI tool made them sound that way. Will we have AI perform our conversations too? Wear a hood and hold up a phone to talk for us?

And the more the AI tool gets used, the more it will shape our tastes, so we like its pabulum more and more. (Sure… there will be edgy AI created stories for outsiders like us…)

Tastemakers in previous centuries could be overthrown, rendered obsolete. Will an evolving tool that lets anyone sound as erudite or as ignorant as required ever let go? Will anyone be allowed to see beyond its anodyne edges?

I don’t think so.

Not because “it” will have agency, but because we (as a population) will have become so conditioned by its products that to break out will seem like heresy. Human genius will be crushed rebellion.

A creative face endlessly stamped upon by an uncaring AI boot, their words, their voice, lost in mangled bone and blood.

This is the slippery slope I see before us, presented by AI tools to writing, to art, to music.

Writing faster, or more, is not better.

Letting everyone speak does not count if they do not use their own voice. And a prompt is not a voice.

But hey,

Why bother working out how words fit together when ChatGPT74 does it for you, and always has?

Why bother thinking at all?