Neighbourhood radical: A story about "AI"

Last night, at a kind neighbour's gentle garden party, I found myself calling for revolution and telling someone to choose which side they'd be on.

I guess that's who I am now?

The conversation was friendly, and we'll talk genially the next time we meet, but it was a strange way to talk to someone I'd never met before who had received the same broad invitation I'd received. Maybe this new friend will be my enemy when the revolution comes. We'll see.

Anyway.

In olden times, when someone heard that you were an English teacher or English professor, everyone wanted to talk about grammar. Usually, someone would open with a variation on "English? Oh, I was so bad at grammar in high school." That's an impoverished way to think of English, and even of grammar, but I get it. English teachers have been complicit in this impoverishment, and some of us have a great deal to answer for. I don't blame these former students for reacting this way, because most of them are speaking out of genuine, lasting trauma that came at the hands (wittingly or un-) of their long-ago English teachers.

Library of Congress, open-access

But it seems that we're over grammar now, mostly. These days, they want to talk about AI.

This particular person said that their daughter teaches English at a community college, and is struggling with all kinds of things related to AI: cheating, originality, form, literacy. He went on to say that in his view, the hard thing is that although right now AI (by which he means LLM-style AI) makes things up, they'll be vastly better in a few years, and then where will his daughter be?

I disagreed. I told him that this type of AI won't get better, not markedly so, because fundamentally it's only a token-suggestion machine, just a more energy-intensive, math-forward version of his phone's text-suggestion tool. It's just that we're easily fooled, as Baldur Bjarnason has explained so well.

ChatGPT, for example, I explained, is stuck with its basic architecture. There's no path through ever more advanced text-suggestion tools to some transformed intelligence, just different versions of what's already here. Its future versions will yield more polished suggestions, ones that look more plausible, but it represents no possible path to genuine intelligence (as Oisin McGann has argued persuasively). As a result, everything will continue to depend on human expertise. Nothing generated by LLM-style AI can be allowed to stand without evaluation by an actual intelligence, and as I think Leah Reich might say, that'll never change as long as we keep trying to build/maintain the right kind of society. Weirdly, even Fast Company seems to be on board with this idea.

(Yes, yes, machine learning, looking for tumours and genetic markers etc. That's not what we're talking about, and you know that.)

His daughter, I said, would be fine as long as her institution stands firm, and as long as we remember as a society that we need lots of smart people so we can be smart collectively.

(No, you're wrong, I really don't mean that we all need to take English classes. Nor do I mean that you can only be smart once you've taken English classes. I mean that universities and colleges are in the business of helping students toward expertise. It's not the only such path, but it's the path his daughter has found herself walking, and this was an actual, personal conversation.)

He asked whether I expected, though, that we're en route to a kind of tech-divided world, with machines and those guiding them (and being guided by them) sitting atop all the money, and everyone else doing what they can to get by.

To my surprise, I just said, "No, because that's what revolutions are for. We can't allow that to happen, and we're all going to have to decide which side we're on."

He was taken aback, of course, because who wouldn't be.

Frankly, I was taken aback as well. I'm not sure I realized how intensely my sense of social obligation has become armoured, over these last years of faux-AI foolishness.

He then went on to talk about a friend of his whose business is the making of dental molds (link only for info: not his firm). This friend had always employed a small staff, people who would stay for a fairly long period and then move elsewhere once they had acquired enough of the right expertise to work independently. Now, though, this friend uses AI to work entirely on his own, producing more dental molds than ever before, each one selling for at least as much as before, with zero labour costs.

How, my new friend wondered, did I feel about that, because you can't stop progress. The dental gent's new waterfront home is lovely, too, my new friend remarked, at which point I realized that it's the monstrosity squatting atop what was once the childhood home of one of my own friends.

That's not progress, I said.

That's unjust enrichment.

That's the pillaging of society through the exploitation of unspoken historic connections between payment and labour, workers and managers, in order to earn too much income and hoard too much wealth. (I paraphrase here. This was all a lot for me to take in, even though I was the one saying it.)

"That's a man who needs to understand that he'll be on the wrong side when the revolution comes," I said, in almost exactly those words.

We could have talked about universal basic income at this point. After all, if our social structures are going to narrow down our options for how to make the money necessary to survive and thrive, we can't lay that responsibility on the individuals being punished by the collective social decision. If we all need money to survive and thrive, and society collectively is going to decide that there aren't going to be enough ways for enough people to make the necessary money, then we need to share money more equitably. Maybe we would've found some common ground, but we didn't talk about that.

We could have talked about personal merit and opportunity, or taxation, or breaking up monopolies, or social planning through legislation and regulation, or regulatory capture, but we didn't talk about any of those things either.

I just said, "We're going to have to decide what kind of society we want, and we're going to have to choose sides. I guess I know which side your friend would be on."

And then we blinked at each other for a bit, and then I went home, because I had a curry bubbling away on the stove that needed tending.

And I'm still thinking about it.

Stay safe, friends. Fuck AI.