I feel like sometimes people get into this sense of like “team based philosophy” or “team based values” where there’s an agreed upon stance to take or thing to oppose and if you’re on the same team then any argument you make against the thing that is opposed is good and valid because it’s a shot fired against the enemy. So there is very little room to be internally critical of the arguments being made by your own team because then you’re not being a team player. It’s not really about making compelling and correct arguments against something so much as coming up with every possible point against it in order to overwhelm someone into conceding without having the energy to interrogate every single point made.
I see this a lot with AI discourse where there are very real criticisms to make of the technology as it is being implemented and marketed by the bourgeoisie and how there are imperialist exploitative labor practices powering the development of the technology, the literal fueling of the technology which is very energy intensive, and in what ends the technology is being used to achieve.
All of these are very condemnatory on their own let alone together.
But what I see a lot more of in the discourse is these very abstract and ephemeral arguments that are very focused on how the machine lacks a soul, or how the present state of the AI output isn’t very high quality, or how the training datasets are somehow a form of stealing or plagiarism (an argument that is particularly absurd with visual generative AI like DALL-E and Midjourney where it is technologically impossible to recreate a given piece of art in the training set and the thing being copied or stolen is so abstract and effervescent that it basically becomes an argument against being inspired by other artists or becomes an argument that only art produced in pure isolation of the world is ethical).
But anyone who tries to disentangle the very serious economic, environmental, and political critiques of AI from the very abstract philosophical objections to AI gets labeled as not a team player and “pro-AI” which gets a lot of horrible vitriol directed at them and a lot of lumping in with the NFT crowd.
Confirmation bias can be a dangerous thing. Agreement with any argument for your team can lead to some pretty dangerous implications. Laws protecting “artistic style” or “writing style” as private intellectual property would make the world objectively worse is a way that could only benefit the same corporations seeking to utilize AI technology to extract value. Conceding that art generated by AI is “fake art produced by a soulless machine” concedes to those corporations that “the AI made it” and not a human using a computer program as a tool, which allows them to circumvent union contracts, devalue human labor, bolster IP claims for corporations, and removes culpability for actions taken via algorithms.
There isn’t really huge philosophical difference that can be drawn between Midjourney and Photoshop, especially since Adobe has been a major developer of algorithm-driven “smart” creative tools since long before the AI marketing craze. It’s always still a human being using a machine to achieve an end. That the algorithm is seemingly being “creative” seems to evoke a moral outrage in people due to the sanctity of “artistry” but it’s still fundamentally the same technological process described by Karl Marx in Capital. The development of a machine which makes human labor more efficient and thus allows for a greater extraction of surplus labor value from the proletariat over the same length of time and thus the same wages.
The proletarianization of artisan classes was a major topic of Capital. Making shoes was once something you commissioned an artisan to do. Now shoes are mass produced by factories. The problem here is capitalism, not that the mass produced shoes “lack soul” or were “made by a machine.” The machine is still run by human labor. The machine is still worthless without the proletariat. As more and more of the labor is mechanized and automated, the amount of labor that is done by the human just becomes even more efficient. Even more surplus labor value is produced over the same amount of time. The value still comes from the labor of a human being. The Bourgeoisie want to claim that the machine made the shoes and the shoes belong to the company and they owe the worker almost nothing. But the machine couldn’t run without the workers. Workers are still involved at every step.
Fossil fuel powered mass production machines have also been absolutely catastrophic for the environment and are incredibly energy intensive. So many unsold shoes end up in landfills. The colonization of Southeast Asia for Rubber and Latex was and has been violent and gruesome and devastating. These factories are often sweatshops.
The parallels are striking between AI and factories because they ultimately are the same economic phenomena under capitalism and require the same political reformations. The means of production must be controlled by labor, and not capital. The machines must be turned toward serving the interests of human wellness and the living ecosystem of the planet—not the interests of Capital. Unbounded infinite growth remains unsustainable. Degrowth becomes more and more of a harsh necessity the more we see unsustainable and destructive growth.
But were sustainable ecosocialist global equilibrium achieved, there is nothing about the technology of generative algorithms which is inherently morally evil or philosophically objectionable, from what I can see. Only insomuch as the running of these computer programs must not throw off the ecological equilibrium with humanity and must not be utilized for dangerous ends, such as exploiting labor and facilitating the generating and spreading of misinformation.
But of course the world we live in at present is not an ecosocialist equilibrium but global neoliberal capitalism. So algorithmic intelligence has only thus far been 95% a negative force for society and the environment and like 5% a source of neutral amusement.
I just think it’s important to retain a focus on the specifics of what is dangerous about this technology than to allow ourselves to fall into discursive traps which are either easily dismissed and argued against by AI salesmen which makes all the legitimate concerns look unserious—or which actually in trying to argue against the use of AI actually reinforce the exact philosophical justifications used by the bourgeoisie to extract more labor value from workers.
H/t txttletale on tumblr who has made some of these same points before in various blog posts.