AI, Luddites, and the Future – Part 1

Artificial intelligence is here to stay. A lot of people are very uncomfortable with that fact. They are uncomfortable to the point of denying it altogether. But their denialism doesn’t change anything. AI is here to stay.

There are four main reasons the anti-AI give for being anti-AI. One is that it creates “slop”. Two is that it plagiarizes. Three is that it is bad for the environment. And four is that it will take our jobs. None of them are particularly good reasons for being anti-AI.

The “slop” argument is patently false at this point. It’s basically saying that we shouldn’t use AI because AI output is crap. It’s not perfect, of course, but that doesn’t make it crap. Even when I wrote these pieces three years ago, Some Thoughts on ChatGPT – Nutmegger Daily, Some More on ChatGPT – Nutmegger Daily, A Third Post on ChatGPT – Nutmegger Daily, I didn’t think the output was slop. I thought it lacked style. That’s a big difference.

In the past three years, AI has improved a lot. I asked Copilot, Microsoft’s AI “companion”, to write me a sonnet. This is the result:

🌙 A Sonnet for a Quiet Afternoon

When winter sunlight drifts through silent air,
It paints the world in soft, forgiving gold.
The weary heart finds gentleness still there,
A warmth that lingers though the day grows old.

In moments paused between the tick of time,
We learn the art of breathing without haste.
The smallest hush becomes a subtle rhyme,
A whispered promise nothing goes to waste.

So let the fading year rest on your mind,
Not heavy, but a lantern in the dusk.
For even in the cold, new hopes unwind,
And life grows sweet again from roots robust.

Hold close the quiet gifts the seasons send—
Each ending folds into a start again.

It might not be Shakespeare, but it certainly isn’t slop. It’s a lot better than I could do with a first draft. I imagine it’s a lot better than most of us could do with a first draft.

The plagiarism charge is also patently false. Anti-AI folks get upset that the AIs learn by trolling all the content on the internet. That means it might have used copyrighted material to learn without getting permission. But I say, “So what?” That’s how we all learn. And plagiarism isn’t about how we learn, it’s about the content we produce. AI is only a plagiarist if the output is relevantly similar to an existing work. I’m not aware of anyone producing an example of that.

The environmental anti-AI argument does have some merit. The server farms used to run AI are very resource intensive. They use a lot of electricity and water. That’s a bad thing. However, it is true of all server farms, AI or crypto or something else. It’s not something inherent in AI, it’s just the state of the computing technology right now. It doesn’t mean we have to abandon AI. It means we have to push for cleaner ways of running it. I’ll come back to this in part 2.

Finally, there is the argument that AI will take our jobs. It is entirely possible that AI will take our jobs. The Luddites, after all this time, may be right. Unlike past examples of technology displacing workers, AI might not create new jobs for people to do. The thing is, losing our jobs doesn’t have to be a bad thing. John Danaher argues in his book Automation and Utopia that work is bad and automating it away is a step towards utopia. We just need a little political and social creativity to get us there. Instead of protecting outdated jobs by suppressing promising technology, we should start advocating for things like a universal basic income.

So, none of the anti-AI arguments are convincing. AI output is neither slop nor is it plagiarism. The environmental concerns are real, but banning AI isn’t the right way to fix those concerns. And losing jobs doesn’t have to be a bad thing. In part 2, I’ll look forward and offer some suggestions for how we should be handling AI.