“There is no alternative” means “STOP TRYING TO THINK OF AN ALTERNATIVE.”

A while back someone contacted The Local Reporter about using AI to help reporters like me cover meetings. They have some LLM that will watch the meeting, pinpoint all the key moments and cut them out so I don’t have to watch them myself. Won’t that save time?

A smart manager friend of mine said large language models/AI are actually good at this. And when we do Zoom meetings for the Wandering Grove anthologies, we’ve used Zoom’s chatbot to take notes. But tracking a meeting or a business conference is different from following a government meeting. Will the LLM highlight the things I’d catch? Will it identify all the points that need to be in a good article? The only way to know is to watch the meeting myself and compare, which erases any gains. So no

Plus I suspect this free service is only the beginning; at some point it’ll be “well now that we cover the meetings, why not have us write them up?” Maybe not but a recent article asserts LLMs are doing exactly that, writing stories for harried editor/journalists who may be the only employee at their current paper. It’s efficient (according to the article) and as long as the editor fact-checks, completely harmless. Besides, writing takes money for journalism and time and papers have less of either so why not? It’s inevitable!

Call me cynical but I suspect a lot of editors who go this route are not going to fact-check. They may not even think it’s necessary — after all you can trust the machine, right? You can’t, as the article notes, but once you start using this kind of shortcut more shortcuts will probably follow. As C.S. Lewis says, the slide into bad behavior — in this case, unprofessional and lacking in devotion to their craft — starts small, then accelerates.

And as the title of this post implies (quoted from the Cory Doctorow piece I link to below) this shit is not inevitable. The swelling use of AI isn’t some natural force. It’s not like cars, telephones or electric power where people rush to use it because it’s so much better than other options, it’s because the tech industry spends billions to promote it as wonderful, awesome, the best thing ever! Everyone who opposes it is the maker of the horse-drawn carriage insisting the automobile is a fad! No, there are ton of serious questions and problems, as covered in some of my past Science-tagged posts. The author of that article runs a consulting firm advising the media on the use of LLMs — I’d say it’s in their interest to see the media use more of them.

As Cory Doctorow points out, the kind of thing that the article is talking about is reducing writers and editors to reverse centaurs. Like the classic centaur, they’re a hybrid of man and machine but in this situation all the power and control lies with the machine. The editor or writer relying on AI has given the machine the challenging, interesting part of the job; all that’s left is checking to see if they’ve done it right. And if a screw-up slips by, well clearly that’s not the machine’s fault — it’s the fault of the human in the loop.

The point of massively pushing AI is not to make things better; it won’t. The point is that companies can fire most of their staff (in a given field), leave the rest as reverse-centaurs and use the labor savings to increase profits. Quality will go way down, but who gives a crap? The salespeople swore it would be just as good! And as Doctorow and Daniel Graeber say, the people who get fired will be capable, highly paid and doing real work — the administrators will, of course, exempt themselves from being replaced, however useless their jobs.

AI spreads partly because it promises a frictionless experience — none of the awkwardness of dealing with flesh-and-blood people and their errors which can be a seductive appeal. At the link, Jill Filopovic discusses why that’s wrong; I’ll add that the idea (which Filopovic quotes, but does not support) that an algorithm, because it’s logical and has fewer random errors, must do better than a human being, is bullshit. So, I think, is Pete Hegseth’s claim chatbots are the future of warfare.

(of course it also spreads because it can do things like make us pay more for groceries)

Of course AI will create new forms of friction, like having to wade through AI-generated bullshit to find a real story or a real author. Here’s an extreme example. Plus accounting departments are seeing a rise in employees who create fake expense reports with LLMs.

“There’s a difference between tools and technologies. Tools help us accomplish tasks; technologies reshape the very environments in which we think, work, and relate. As philosopher Peter Hershock observes, we don’t merely use technologies; we participate in them. With tools, we retain agency—we can choose when and how to use them. With technologies, the choice is subtler: they remake the conditions of choice itself. A pen extends communication without redefining it; social media transformed what we mean by privacy, friendship, even truth. ” — from an article about how higher education is breaking itself by embracing AI and trying to make teaching and learning both frictionless.

Chanda Prescot-Weinstein suggests part of the problem is the increasing emphasis on education as nothing but a tool for our careers and a hoop we have to jump through, and that this is wrong: “Knowledge is worth your time because of how it shapes your mind. And the authoritarians may take many things from us, but they cannot take our minds (unless you let them). So, I know it’s very hard right now but that intellectual work is worth your time, even when it’s not obvious how you will profit from it. You are more than a future source of profit, and humanity’s survival depends on all of us understanding this.”

Against this argument we have dubious claims that AI will democratize education so schools must adapt or die: “A business student can ask ChatGPT to explain supply chain optimization or generate market analysis in seconds. The traditional lecture and test model faces its blockbuster moment.” Because before ChatGPT nobody ever had the option to look things up in a textbook or a library book. Apparently, though, there’s a cottage industry built on promoting AI to students.

Then there’s the perennial appeal of using AI to create your own media. As I’ve said before “give AI a prompt” isn’t creative — and I’m curious how many people are really keen to do this as more than an idle amusement (let alone if, as the article notes, they can’t infringe copyright). At least one videogame company has dismissed this sort of thing as “creating” for people who aren’t creative.

The AI Is Inevitable crowd skip over the price of using it too, like the demands data centers make on power grids and water supplies. That said, the Felon of the United States is on the side of the AI techbros (though that doesn’t stop him selling computer chips to China) and they are determined to fight any restrictions on their delusions of grandeur. Although the collapse of the AI bubble might.