It's a form of productive capital, and its ethicality depends entirely on the conditions and reasons for its use, same as any other. Renting time on a proprietary model is unethical, because that represents a modern enclosure of the cultural commons. Capital replacing skilled workers with generative slop is unethical, because it's yet another step in automating away productive workers with inferior machines.

Hobbyists and yeoman artists using non-proprietary local models to amplify their own labor is ethical, because that's just a worker using a tool to be able to accomplish more, while existing as much as possible outside of the cultural enclosure techbros are trying to make.

The question of whether training data has to be properly licensed from someone who claims the rights to do so is a red herring meant to favor huge corporations that either own massive amounts of IP (like Disney and other media companies) or which claim licensing rights over massive amounts of user-provided IP (like , imgur, instagram, etc) and who can negotiate licensing fees out of the big tech companies like what got from google. The property angle will never come out in favor of small yeoman artists and their meager holdings any more than property rights came out in favor of yeoman farmers over huge agricorps, and thus should be disregarded.

That said, the hobbyist AI community is at least 95% irredeemable and better off in , from the grifters, to the nazis, to the nonces, to the people whose only crime is just being too cringe.

There is no "ethical" evaluation of AI under Marxism. It's clear that AI is merely a tool, like all other forms of automation, to displace workers for the sake of profits. AI isn't "Good" or "Bad" in this context. Under the constraints of capitalism, it will be used for "Bad" things, meaning "non-productive" things. Sold as a toy to users to perform whatever they desire, creating nonsense text and images that ultimately have no value. AI could be a truly transformative technology if it was confined to a more socially responsible system. It's use in protein identification, for example, is a real leap forward.

Also, like so many "revolutionary" technologies, you'll see capitalists bend over backwards to add "AI" to whatever it is they produce. That's how you get bullshit like the "AI" mouse, or the "AI" tooth brush. It is also a smoke screen for ACTUAL intelligence that is being exploited through Capital's imperialist tendencies. Those little coolers that use AI to drive a subway sandwich to your apartment? It's "AI" is probably named José and José gets paid $0.10 an hour to drive that little cooler to your house. LIFT or Uber (I forget which) uses "AI" to identify the driver based on a photograph to ensure people are not "sharing" the account (and thus able to be on the road for longer than a single person could). Again, that "AI" is probably named Isabella, and she was paid $0.05 from a microwork platform, and did the "computations" to decide if today you looked like that photo of took when you signed up for LIFT/Uber.

Like all things under capitalism, these automation tools are used in ways that harm workers. This does not make the underlying algorithms and their many applications inherently "unethical". It's the actions of the capitalists, that are ultimately unethical.

Freelance artists hold the contradictory position of being semi-autonomous wage-earners. I find this diagram helpful for understanding why they tend to hold certain positions. Also explains why they ally with big companies using IP law against LLM companies. Because they don't want to become proletarianized.

So as a force of proletarianization, wouldn't the technology be historically progressive in a Marxist sense? I still hate it though.

Read the Communist Manifesto. It's been a while, but iirc, a significant chunk of it is about Intellectual Property and automation.

I don't see how it's any different than other technology tbh. Most discussions of ethics in this context are committing some sort of scope error, where the implication is that any one individual's choices in this regard have a meaningful impact. Either that or we're talking about some fantasy where the working class has any amount of class consciousness and is able to act as an entity in its own class interest. I won't fault anyone for avoiding unpleasant vibes, but on an individual level none of that is particularly Marxist.

Wrt the analysis, in addition to the great points already mentioned in this thread:

The explosion of the internet created a new sort of frontier with untapped resources and unenclosed commons. Billions of people passively and actively creating art and information and data points for decades, most of which freely given or taken by tech corporations. Part of the trick is that this stuff wasn't really a resource at the time, not in the sense that it is now. The question "how can someone own my conversations, my habits, my preferences and tendencies and opinions and thoughts" mirrors "how can someone own land?". A social transformation, a dialectical development, a de- and reterritorialization.

AI models require a tremendous amount of data to train. One of the LLM models, for example, needs about 70 years of input in order to learn a new language (as an aside, compare that to the ~1500 hours it might take a human). The end result is an incredibly useful and valuable machine, capital, imbued with a tremendous amount of dead labor.

The high barrier for entry means a further concentration of capital in every industry where AI can effectively be utilized (and isn't just a gimmick). The AI-owning bourgeoisie are incentivized to heighten this barrier for entry, and this will happen in lockstep with how much doing so decreases their flow of cheap data compared to how much data they need. But the energy and tech cost is already high.

This abstract notion of "data value" is transforming into a concrete one, and with that comes the enclosure that's characteristic of a property economy.

Now, this could be a lot more impactful than many online leftists seem willing to admit, but it's still taking place in a highly abstracted place with tenuous ties to the material mechanisms and primary contradictions of society. Part of the difficulty in analyzing it comes from the spectacle nature of these abstract realms; they can replicate or imitate similar processes that happen in the material world, like enclosure and exploitation, but that doesn't reveal what their mechanism in the actual material world is. Does the existence of an AI that can speak French change the flow of resources from Africa to Europe?

I think its impacts will closely align with the scopes it exists in, so more impactful within the abstractions it operates in, like the internet and media, than base material flows, where it's mostly just a resource sink. From an international perspective, it's more like a reorganization of the lord's manor than an actual restructuring of the system at large. A change in how spoils of empire get divvied up is largely immaterial, in this scope.

As it consumes its abstract frontier, though, many people that subsist there might find themselves proletarianized (or "materialized", forced out of the digital proletariat into the manual proletariat, echoing the historic flow from countryside to city). That's where I find myself personally, with the work I've done for 10 years quickly disappearing.

1. Probably not, I'm not gonna complain if people use ChatGPT. But maybe someone has more knowledge than I do. I'd try not to pay for AI, if possible. If you're a company and you're profiting off of AI, then you're stealing the work that was used to try AI.
2. Do you mean as an engineer? Sometimes giving up work is hard.
3. You don't mean "all" that comes up in the URL?