reading and writing and ai

new tech and the framing of luddites

In the late 1980s, John Fogerty, writer of Proud Mary and the man responsible for the entire back catalogue of Creedence songs went to court to prove that two of his songs sounded similar because, well, he wrote them both.

He took his guitar into the courtroom, and convinced the jurors. The recording company Fantasy, that owned his music, had accused him of copyright violation of his own song (because at the time they owned the copyright to all of his music.) They lost.

Fast forward to the 21st century, and we have ai taking the work of musicians, writers, artists and other creatives to produce ai-generated songs, books and visual art. Ai-generated songs are reaching No 1 on various charts (in the USA country, blues and R&B).

Substack has not been immune to this artificial ecosystem, becoming problematic for Ai-generated notes and posts.

On the weekend that's just passed, Substack introduced integrated ai-detection software.

Naturally, as this is a writer’s platform, a whole lot of writing emerged on this issue.

Many writers pointed out their pre-ai novels and theses had been stolen and fed into llms, and now the detection machines were branding their own work as ai (as the ai had fed on their previous writings.)

Andrea Bartz , one of the authors that took Anthropic to court, forcing them to make copyright compensation for stolen novels and books, describes the process of writing her novel then finding her work gobbled up by llms here.

In the discussion that followed on posts and notes, a common point was the ridiculousness of using a machine to detect machine writing. Many people pointed out that we already have a slop-detector machine, it’s called reading.

Others — and I’m summarizing what I’ve read here on Substack, not commenting on the points — wrote about the unreliableness of ai-detection software, the manner it which it determines ‘human writing’ and how it is discriminatory on race and neuro-function.

The debate made me recall the Luddites - a group of people who were framed for being anti-tech. More on them later.

At a time when authenticity and creativity are suspect, transparency is crucial. Thus, how I engage with Substack is as follows.

My approach to reading and writing here are two different things.

Reading

I have two criteria:

Do I perceive this writing to be quality?

Is the use, or not, of ai transparent?

I don’t care if you use ai to help you structure you work or edit. If it’s interesting and well- written, I’ll read it; but the minute I read too many short, quick, its not this it’s that statements, I’m out. I just can’t do it. At the very least, writing should be varied. I love reading human-written work, as they are usually very nuanced and original, containing combinations of words I find beautiful or interesting.

Writing

I do not use large-language models (ai, llms) in any shape or form.

No-one ever writes in isolation, so my writing and thinking stem from what I have personally read, conversed about, discussed, read again, and thought about.

I do not use ai to give me ideas, structure my article, do research, summarise research, nor for grammatical suggestions.

I write, read what I have written, edit, reread, edit, restructure myself.

When I research, I copy, paste, highlight, reread, think and write a summary.

Personally, I do not wish to use a technology which has been developed off whole-scale theft of others’ work, and has implications for the future of education, creativity, and even, it seems, human cognitive function.

However, I accept that we all do not have the same ethical approach to ai, and that is why my reading guidelines are different from my writing guidelines.

Back to the Luddites.

Luddities, named after the mythical King Ludd of Sherwood Forest (yes, that Sherwood forest) have gone down in history as being backward opposers of any form of new technology.

But the Luddites, as it happened, were framed by the 19th century version of mainstream media. They were caricatured as being anti-tech, when in actual fact, they were pro- workers rights.

They crept into a factory, one dark dark night, and destroyed a whole lot of frames. Stocking frames, to be exact. Stocking frames that produced low quality goods, that used to require intensive labour from skilled framework knitters.

The new mechanical frames were on looms, and the steam-powered frames could do the work of multiple knitters. So a group of knitters waltzed into a town in Nottingham, five miles north of Sherwood Forest and destroyed a few looms. Sixty-three, actually.

They didn’t wantonly destroy the new tech – they targeted factories with very poor industrial relations.

The ones with low pay, the employers who broke workplace agreements, the managers who provided poor working conditions.

The Luddites formed workers alliances and lobbied for quality control of fabric. They protested against exorbitant price rises for cloth.

So what did the industrialists and the British crown do?

They hired a militia of course, goons paid to kill the protestors and stacked the courts to deport or murder Luddites.

The word ‘luddite’ has come to be used in a derogatory sense, someone who is opposed to any new technology.

Cambridge dictionary says:

Luddite: a person who is opposed to the introduction of new working methods, especially new machines

The Cambridge dictionary is wrong. When Netty and Ned Ludd and their band of merry people wanted to engage with the powers that be, they wrote letters to the governing officials. They all signed them, without fail, King Ludd of Sherwood Forest. Ned Ludd may or may not be apocryphal, no-one really knows.

Regardless, the Luddites used Sherwood Forest as a brand name, a place that everyone in 19th century Britain knew was code for advocating for economic equality.

When they were destroying the new tech of the day (automated looms), it wasn’t the tech per se they opposed. They opposed the owners of companies that tried to limit worker’s rights, lower their wages, make working conditions hell, and increase prices. If the Luddites were around today, they’d be targeting their operations. They’d probably start with palantir.

The Palantir that has an agreement with the Israeli Ministry of Defence. The Palantir that allegedly uses ai to target people in Gaza.

So whilst it’s crucial to be transparent about our use of llms and other ai, and to consider the deep human and ecological consequences of these technologies, we should remember there are people who are paying in unimaginable ways for the rapid expansion of ‘artificial intelligence’ with the military-industrial complex.

We live in strange, difficult, and deeply confronting times. Within the Substack ecosystem, we’re making it up as we go along, reacting when we are presented with business decisions, and showing that if engage in debate and display our concerns, we can effect change.

After the intense reaction to the introduction to an ai-detection integrated software, Substack changed their delivery approach. Everyone who posts now has the ability to block the ai-detection software on that post. For the moment — unless enough people clamour for further change — it’s a post-by-post function. Many are advocating for this function to be publication-wide.

I really value the people on Substack: the writers who write interesting, unique pieces, the journalists who engage in independent journalism, the truth-tellers who remind us what’s going on in the world, the note-posters who alert me to something more deeper or create a funny paragraph that makes me laugh out loud. All of the people I ‘ve chatted with through the message function, the poets who build magic with line-breaks and rhymes, the philosophers who muse over the unknown, the scientists who break down difficult concepts, the travel bloggers who bring the world alive, the historians and colour-pigmenters and rock-analysts and tea connoisseurs, everyone, and you, dear reader, you, are what makes Substack what it is.

Let’s keep reading and writing and speaking up loudly about things we do not like here, so we can make the platform better.

Perhaps enshittification is inevitable, but perhaps, just perhaps, englorification is possible too, if we all just keep trying. There’s lots in the world that’s glorious. Here’s two:

The beauty of living on a peninsula is that sunset (left) and sunrise (right) are both within walking distance of my home.