Not Generated There

The Divine Sarah points to this article, which is about using A.I. to write something coherent instead of:

Often I’ll write a paragraph and think, “This sentence is terrible.” Not grammatically terrible, structurally terrible. Like it wandered into traffic and got hit by two semicolons.

So I’ll ask AI to smooth it out. Sometimes I use its suggestion. Sometimes I hate its suggestion. Sometimes its suggestion gives me an idea that turns into something completely different.

It’s less “robot writes article” and more “robot sits in the corner saying, ‘Have you considered not writing that sentence like you’re being chased by bees?’”

That’s useful.

There have also been articles that were difficult for me to write.

You know the kind. The ones where your brain is running at about eighty miles an hour while your fingers are trying to steer through tears and frustration, and what comes out looks less like prose than a transcript of someone arguing with a smoke detector.

Those articles were mine.

Yeah, well.  Now I have a confession to make.

I’ve never used A.I. to wrote anything — hell, I don’t even know how to access the thing.  Everything that you have read on these pages for the past twenty-off years was conceived, written and edited by me.  The only time I’ve resorted to technology is when I’ve used Spelchek to make sure that in my haste to get something down, I’ve put “i” before “e” all over the place.  But to use something to check my style or flow?  That’s part of writing, and that’s what I do.

That’s not going to change.

Don’t get me wrong;  I’ll often write something, then when re-reading it, think, “What the hell have you done here?” and set about fixing it.  It’s time-consuming and often exhausting, but it has to be done.

Of course it’s easier to use A.I. to help you write something.  Those of you who have followed my train of thought on the topic, however, will know full well the contempt I feel for using technology to make one’s life “easier”, at the expense of losing control and (perhaps) becoming part of the Machine/Hivemind in the process.

Also, I think that using A.I. is cheating.  When you read something written by someone, that reading should give you an insight not only into the topic, but also into the mind of the writer.  Otherwise, you might as well just give your thoughts to a panel of writers, speechwriters and PR people and say “Here.  Do something with this,” and then publish it under your name.

Which is kinda what you’re doing when you get A.I. to write something for you.

And if you can’t write something without this:

Often I’ll write a paragraph and think, “This sentence is terrible.” Not grammatically terrible, structurally terrible. Like it wandered into traffic and got hit by two semicolons.

…then you need to learn how to write and edit your own stuff.  It’s hard — really hard, ask me how I know this — and often you’ll fail.  In which case it’s time for a rewrite, or a visit to the trashcan.

Mark Twain used to read his daily output to his wife and daughters, because he found that not only could they spot problems in his writing, but the very act of reading something out loud highlighted mistakes and missteps.  He was lucky;  most of us don’t have that sounding-board, and we have to rely on our own devices, skills and indeed patience to get the damn thing right — which is what I do.

And at the end of it all, when that piece of writing is finished, I can look at it with if not pride then at least satisfaction, and say:  “Look wot I done.”

Which is what makes it all worthwhile.  Writing is not a collaborative activity:  PR releases and corporate statements are.

Thanks, but I’ll stick to doing it all on my own.  My Loyal Readers deserve no less.

A.I. delenda est.