Smashing Frames

With studies upon studies showing that actual measurable productivity gains through “AI” (which these days basically means chatbots) are really hard to come by and that “workslop” (meaning the extra work created for the rest of the organization by one person using “AI” lowering their work quality) eats up a lot of what might have been productivity gains many critics feel somewhat relieved: When the main narrative of “AI” (massively increased productivity) fails, this surely marks the end of this bubble and we might be able to get back to talking about actual problems.

I am not so sure.

Not because I think that some tweak to some LLM will make them suddenly that much more reliable or useful but because it was never about productivity really.

It’s something I alluded to in my “AI” talk at 2023’s Re:Publica: It doesn’t matter how good these systems are in reality, because that’s not what your boss cares about.

“AI” is a tool to disenfranchise labor.

That’s the job. If “AI” is actually more expensive than paying actual people actual wages that’s still a good investment for capital because it is about breaking up the structures, networks and organizations that help workers organize and fight for labor standards and fairer wages.

When CoCa Cola creates another bad ad using “AI” the fact that it’s garbage and expensive isn’t the point. It’s all just an investment into no longer needing to pay people for their expertise, work and time.

In December 2024 Ali Alkhatib wrote:

“I think we should shed the idea that AI is a technological artifact with political features and recognize it as a political artifact through and through. AI is an ideological project to shift authority and autonomy away from individuals, towards centralized structures of power.”

This is the macro-level view. It’s about Google, Microsoft and Amazon digging themselves even deeper into not only your personal life but also every economic workflow and process in order to charge rent using the established dependency.

But on a smaller (a bit short-sighted) level CEOs are looking towards “AI” not so much as actual replacement for labor but as the leverage to push down cost in the long term by crushing labor power.

So showing that productivity doesn’t measurably increase is winning the wrong game. It’s not an attack on our capabilities, skills or experience, it’s an attack on our collective and individual power.

Liked it? Take a second to support tante on Patreon!

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.

“AI” exists to disenfranchise labor. That’s what it’s for. Regardless of how good these stochastic systems are or the flaws they have just being able to point at the non-unionized robot whenever the employees ask for raises or anything really is incredibly valuable for business. The existence of “AI” and the supporting narrative mean that you defending your value and therefore price on the market will have a way harder time.

But that’s the impact for people on the top. That’s why the C-Suite wants those tools to exist and to be deployed. I think there is also a strong effect on the way “AI” affects the social relationships between peers and coworkers.

Work is not really about finding friends. It happens but it’s not necessary. But there is – ideally – a level of respect for one another. The acknowledgement of each other’s expertise, skills and contributions as well as one’s weaknesses. All those allow the forming of communities based on solidarity: If the people touching computers realize that the workers in the warehouse are part of the same struggle all those groups together can actually develop some power to make their needs heard and acted upon. If for example a company manages to separate the different groups and departments, plots them against each other, this weakens everyone but management.

And I think “AI” has this massive effect on feeling connected with each other, in seeing each other as peers.

Let me tell you a story. A few weeks ago a colleague was talking about an issue he had with a specific piece of software that we are – as we usually do – using a bit outside of what it is intended for. He outlined the problem and then went into the things he tried to solve it culminating in an explanation of what probably will work. Another colleague I respect a lot then responded with “did you ask the AI?” and it felt like a scene from a movie where the protagonist is just sitting there, hearing something and suddenly a big dude comes in and slaps him in the face with a fish. I was irritated and it really took me a while to understand why.

It was not the absurdity of the statement (What even is “the AI”?) or the weird dynamic of asking a colleague who just presented his solution with the suggestion to use an unreliable search engine instead. It was me feeling drifting away from that person. Because that statement made it so obvious how every “here’s a problem” is now connected to “let’s ask the spicy autocomplete” for that person.

In the weeks after that whenever there was an issue with the stuff coming from that team, when something wasn’t up to par, or had a weird structure I realized how I kept attributing it to them leaning heavily on slop machines. Which isn’t necessarily true: Shit happens, software and hardware is complicated and sometimes things end up weird or hackish or broken – regardless of what tech you use.

But I realized that this event (and all the little similar events before) ate away from a trust relationship that had been developed over years. I was having sort of the perceived “workslop” experience.

“Workslop” is a term that defines bad, “AI” generated work product that someone produces to fulfill their work duties on a surface level that their coworkers will have to clean up: I generate a bunch of code that kinda works and someone else realizes that it’s a hot mess when trying to run it and has to clean up the mess. In that example I would have produced workslop (but might have been very efficient!).

The experience of workslop (whether it’s real or mostly perceived as I showed in my short story) directly erodes social connection, erodes trust in one another and in the end erodes solidarity. Because why would you stand with a person who does not do their job and offloads their work on you?

Pride in one’s work under capitalism is a bit of a weird thing: You can be proud of what you did but that great work will more often than not not be especially rewarded. But what it does is signal something to your coworkers: Doing good work means that you respect the people working with you, working off of what you did. It means that you try to make their lives as easy as possible because you’re all in the same boat.

“AI” pushes everyone towards slop. “Just do it. It’s easy. You look innovative. It’s fast. The others can also just use the slop machine to fix the mess if it occurs.” But the dissolving effects on the social fabric of the workplace cannot be underestimated: We are already working in conditions that make building the foundations of solidarity harder. Teams get pitted against each other based on KPIs, everyone is working from home, never having to meet their coworkers. As usual “AI” is just gasoline on the fire. But maybe we should start putting out some fires?

Liked it? Take a second to support tante on Patreon!

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.