Frigates and ChatGPT

How to Get Away with Murder No Foreign Sky

Frigates and ChatGPT

This post on Writer Unboxed caught my eye: On Frigates and ChatGPT

This post starts with a poem by Emily Dickinson:

There is no Frigate like a Book
To take us Lands away
Nor any Coursers like a Page
Of prancing Poetry –
This Traverse may the poorest take
Without oppress of Toll –
How frugal is the Chariot
That bears the Human Soul –

I’ve always particularly liked Emily Dickinson, and this poem is so appropriate when thinking about fiction. That’s exactly what fiction is supposed to do: take us lands away. Also, just offhand, I would agree that this is exactly what fake fiction generated by AI programs has no hope of doing. I presume that’s what the linked post is going to argue, and I don’t think it needs argument. I think it’s plainly impossible for a non-conscious text generator to create fiction that is thematically coherent and meaningful, never mind coherent in terms of plot and characterization.

Let’s see what the post is actually saying, however:

How do you see the world? This is the tacit question we ask an author each time we slip into a story. Reading is the closest we can be to understanding another person’s way of seeing. We are all trapped within our own consciousness. We peer through frames we are often unaware of such as culture and language, family and upbringing. Even the physiology of our eyes plays a role in determining how near or far we can see, as well as the colors and shapes of the world around us.

Yes, well, differences in sensory experience are one thing, and differences in the experience of imagination are another, and those differences can be important and interesting. But surely what we mean here by “How do you see the world?” is “How do you see the world philosophically?” Or even, “How do you see the world morally?”

More than that, fiction also asks, “How would this other person see the world? This person who is not me and also not exactly the author? What would it be like to see the world this way?”

All those questions are obviously meaningless with fake fiction generated by non-conscious text generators. That’s why “artificial intelligence” is a misnomer, and one that I think may not be harmless. The things aren’t any more conscious than a tree, far less a dog. They generate the appearance of intelligence without the least trace of actual intelligence. It’s really odd to think about.

I haven’t yet seen student essays that I can easily tell are ChatGPT-generated, but I really am not concerned that text generators will be able to create coherent stories for a good while yet. I wonder if I could be wrong? I don’t think I can be, at the level of a novel. There’s just too much in a novel. Maybe a really short story where everything is a clever plot twist and there’s no depth? I’d be interested in seeing some of the fake stories being received by short story publishers now.

Meanwhile:

There is no Frigate like a Book
To take us Lands away