Disclaimer: I am currently employed by a company which utilizes machine learning and artificial intelligence. This blog post was written prior to this employment, and in no way reflects the views of my employer.. I’d imagine. All views are my own and all that.

Recently I’ve been watching as a number of online personalities I enjoy have taken their shots at Artificial Intelligence, discussing why the technology is over-hyped, why it’s bad for workers, and the damages it can cause. However, I think their focus on the quality of current AI implementations, and it’s derivative and repetitive tendencies, results in them losing sight of the main externality that they should all be worried about, wealth distribution in an increasingly automated society.

Some examples of the content I’m referencing are:

Adam Conover’s A.I. is B.S., and A.I. and Stochastic Parrots

John Oliver’s Last Week Tonight – Artificial Intelligence

Some More News’ What are the real dangers of AI?

If you don’t feel like listening to 3 hours of content, or are listening to this article, most of these videos / podcasts revolve around a few core arguments against the technology:

1. All output is derivative of the inputs. ChatGPT can’t write good movie scripts, and Midjourney can’t produce “real art” because all it’s doing is re-configuring things from it’s dataset that already exist, and therefore it’s not original or creative, it’s stolen and derivative.
2. The technology isn’t as good as is advertised. Musk promised self-driving cars, and they’re nowhere to be seen. Some company implemented an AI chatbot in a support hotline and it said something horrible that a human never would have. etc. Basically “haha look how bad this is, it’ll never take my job.”
3. AI is a danger to jobs and only serves to make the rich richer. If it is as good as advertised, it will be bad for humanity because it will lead to people losing their jobs and the products of their former workplaces will decline in quality due to point 1/2.

While I won’t address the points 1-by-1 because I do somewhat agree with them, I think the main issue that’s been nagging at me as I’ve digested these arguments, is that by focusing on the quality of the technology, and the processes by which it “creates” (points 1/2), it pushes the conversation towards laws, limitations, bans, sanctions, and bargaining agreements which don’t fundamentally address the wealth that AI stands to create. It aims to inhibit the wealth, rather than to appropriate it.

An example of this is the Writer’s Guild of America’s current strike. One of their main points they aim to address in their bargaining, is the issue of studios turning to AI to write content, and using writers only for editing or “refinement” jobs. The writers want a clause that prevents the use of generative language models to write scripts. It’s understandable. They’re a union, and they want to protect their jobs. However this is one of few places in my personal philosophy where I find myself up against traditional union action. While I support unions in the broad strokes, in my opinion, banning the technology from your industry is a poor way to adapt to changes that disrupt it. It’s analogous to farmers revolting against tractor manufacturers because they’ll need less farmhands to till soil. It’s a losing battle. Writers simply will be less necessary as time goes on. It may take time as AI improves, and it may not be a full replacement. It may be a situation where a writer generates the scenario, and the AI fills in the dialogue and character backstories or something, like this fictional scene from the series “Westworld”.

Whatever it looks like, AI assistance is going to reduce the mundane parts of many professional and knowledge worker’s jobs, and it seems to me like most people’s response to this, is that they want to throw it in the trash so they can keep their job security. They want the tractor to die so they can continue tilling their fields by hand.

Writing was just an example, but I think many industries will look like this in the future, where a single person, assisted by AI, can produce what took entire teams nowadays. When computer animation became popularized, the demand for animators dropped significantly. The 1994 Lion King movie had an animation department of over 400 staff, while 2022’s “Lightyear” had an animation staff of 98. The rapid acceleration of technology to provide easier tools for these specialized tasks has reduced the workload by four times while simultaneously increasing quality, and that’s before AI even enters the picture. How long will it be before single-digit AI assisted animators can provide scratch-paper storyboards, some reference art, and then talk an AI towards a 90% finished product?

We already have a bot that can write functional code from a plain English prompt. While it still requires a bit of know-how to integrate the blocks that it produces, it can, with great efficiency guide beginner to intermediate programmers to solutions that would have taken them hours. This type of assisted work-acceleration is real today, and it will expand to other domains as the technology advances. Text, code and art are the prime candidates right now because they’re the easiest to find mass data sets of online, but as we figure out better ways to capture other disciplines, human labour will become less and less the norm, and human guidance will take precedence. AI does not need to be able to do someone’s job entirely, it just needs to help one person do their job in a fraction of the time, and suddenly that team of 10 becomes a team of 8, or 5, or 2.

I could continue to illustrate the point by pointing out tech innovations that have reduced worker demand, but in the interests of brevity, I’ll return to my main point, which is that legislating away the technological advances is not the key to human prosperity. We, as workers should not be rallying and protesting for the right to continue doing work that has been automated effectively.

We should however, be adamantly protesting for a significant slice… or all of the proceeds of our jobs being automated away. A common refrain from socialists, communists, and anarchists like myself is to “seize the means of production” – to use the collective power of your labour to appropriate the factories, the land, and the machines that produce wealth, and to do exactly what you already do, but for your own benefit, and not for the benefit of an arbitrary owner of this “private property”. However we’ve entered into an age where these owners are moving to, en masse, replace that labour power with digital scabs. Scabs that can’t be swayed by picket lines, solidarity, or bad press. The bargaining power of labour is going to evaporate, and if workers don’t use it while they have it, they’ll be left behind destitute like this excellent scene from “The Expanse” which illustrates the poverty created by earth having a population in the billions with an actual job market that can accommodate millions.

The idea of Universal Basic Income has been gaining traction all around the world, and the biggest question that always gets raised, is “how will it get paid for?” Well here’s an idea, how about with the profits of every job that’s effectively eliminated from the global workforce? Why is it that when a job is eliminated, the company sends that worker away and keeps their pay cheque? Why can’t two workers with the same job just work half as much for the same pay?

Remember when grocery stores used to have 7-15 clerks, with a bagger at each til? Now think about how when you go to a grocery store now, you scan and bag your own groceries while one person stands at a booth keeping an eye on their console for suspiciously heavy bundles of bananas? (4011 for cheap steaks by the way). Do you ever stop to ask what happened to the $200/hr that just evaporated from that store’s payroll? It certainly isn’t making my groceries any cheaper – Galen Weston, you smarmy fuck.

Taken to the absolute extreme, AI could potentially lead us to a post-scarcity society, if we’re not already there. A recent study found 19% of employees in the US felt that their jobs were “socially useless”, and I’d personally argue that far more are just unwilling to admit it. Why are we so insistent that when a fifth of the population could go on holiday and nothing would change, that we must keep finding things for them to do? As Artificial Intelligence becomes more prominent, we’re only going to see this number rise. I would argue that as a society we need to start shifting our priorities. We need to pursue work reduction as a goal.

This is the failing point of capitalism. We have chosen an arbitrary point system (dollars) to represent value, but it is fundamentally detached from human needs. Profiteers can earn points by creating starvation, so starvation exists. Profiteers can earn points by withholding housing, so we have housing crises. Profiteers can earn points by automating work, so wages plummet, and profit soars.

As a collective, we need to keep our eyes on quality of life increases, while reducing hours worked and labour expended. Human’s labour hours have continually decreased throughout history, but for some reason, after a Nazi-sympathizing war profiteer instituted an 8 hour work day to run his factories 24/7 on 3 cycling shifts, we’ve never thought we could do better. As we march towards a future where at the very least, AI will assist workers in producing technology, food, medicine and more, we need to accept a general truth, that the net production of a human worker, is going to soar beyond the modest ratio to the average human’s consumption that we have today. Currently our solution to that, is to invent jobs that do nothing for the sake of everyone being equally miserable, and then to allow the excess points to accumulate in the pockets of a few assholes who think that strange women lying in ponds distributing swords is a basis for a system of government.

We need to step beyond this and imagine a world where a successful automation of someone’s profession is a good thing for everybody. Our systems need to celebrate, accommodate, and distribute that wealth. It can be hard to see that as a possible reality though. As I wrote this article, I even found myself steering away from my own thesis. I wrote an entire section that I’ve now deleted dedicated to Robert Miles’ video on “Windfall Clauses”, where I gave credit and appreciation to the idea that we could pursue shared benefits of AI. However the more I stewed on it, the more I realized that the idea he poses, is grasping at straws within a capitalist system. It’s begging for crumbs at the master’s feet. He suggests that we should simply ask companies nicely to sign an agreement to share the profits of an Artificial General Intelligence breakthrough for the sake of positive press and moral decency. As I sat and reflected and rewrote the part over and over trying to find the right words, I realized I just needed to scrap it. There’s not a solution in a system with the incentives we’ve created. So long as those controlling the technology need to make their lines go up, we will not see a net societal benefit from Artificial Intelligence.

To conclude, I think those fighting the rise of AI need to reassess. The WGA, SAG-AFTRA, the Youtube creators above, and everyone else poised to lose work to AI, needs to keep their eyes on the prize. AI will not be legislated away, and it is coming. Maybe not in the form it takes today, and maybe not without assistance, but AI will have the power to reduce the labour required to produce the majority of your media, your food, your infrastructure, and more. Instead of clamouring for a continuation of the 40 hour work week, integrate the technology, and while you’re still a critical part of the process, leverage the productivity gains to vie for proportional returns to your output. If we allow the ruling class to siphon away more of the productivity gains from the industrial and technological revolutions than they already have, we will be in trouble.

There are two worlds ahead of us. In one, generative AI’s are gated, private property which generate private profits for the already wealthy, eliminating jobs and causing mass swathes of unemployment and poverty for which there is no recompense. In another, workers pursue control of the usage and benefits of these systems, and demand or force a claim to the benefits of their displacement. I think that as workers, we’re fighting a losing battle by fighting for labour rights. We shouldn’t want the right to work. We should want the right to a living.