The one used in this thread: https://lemmygrad.ml/post/12973939
Mind you, I didn’t set out to deceive anyone. It was more of a “why call attention to it” thing. It’s probably the only time I’ve posted anything AI where I didn’t explicitly call attention to it as AI. But I’ve learned in the times I did do that here that some person or other would be upset, pretty much no matter what the content was of what I posted.
I figured, one of the primary complaints about AI generated images, is essentially that they look like they’re trying to look professional but make basic errors. Which I think is a fair criticism. So I thought, what if I use stick figures. That’s closer to meme level of illustration and expectations are lower for it, so if some basic mistakes exist, that’s par for the course.
I did write the text itself as rhetoric and fictional conversation, but I didn’t draw any of the image, nor did I place the text in the image. It was received extremely well for a lemmygrad post. I’m sure it is easily the most upvoted post I’ve ever made. I don’t know what made it work so well, maybe the commonness of issues with landlords, but I can only guess as the responses didn’t tell me much about the why.
One person did say “Ai stickfigures?” which surprised me a bit. I didn’t think it looked overtly AI genned. But I asked them “why do you ask?” because I wanted to be sure whether it was confrontational or sincere interest in where it came from. They never responded, so I’m not sure.
But anyway, the fact that it was AI somehow didn’t prevent it from seemingly resonating with people. I never would have drawn any such image on my own. I could make an attempt, but it would probably look like a child’s fridge drawing and I would have especially struggled with the vividness of the facial expressions, which was where I sought to do the contrast between the feigned victimhood followed by literal mustache twirling villain. Caricature, but gets the point across.
Given how I’ve seen some react on here, I can’t help but think it would have been a shitshow if I’d called attention to it being AI generated or if it had been drawn in such a way that it looked stereotypically “AI”. Instead, as of posting this, it totaled at 111 upvotes and 0 downvotes.
It’s always hard to tell the impact things have in a virtual forum, but with the limited metrics I have to go by, it went really well. And I didn’t draw a pixel of it. This paints a picture for me that the issue with AI generated content in these spaces is not when it’s AI generated but when people believe that it is. But people can believe something is even when it isn’t. Or in this case, likely assume it isn’t when it is. Maybe I’m wrong and the 100+ people who upvoted thought “AI” and upvoted anyway, but given how things usually go here, I doubt it.
I hope no one takes this as trickery or something. Having seen more recent angst on here relating to AI, I thought calling attention to this is an opportunity to make a point. When I went to make that image, it was mainly about seeing what kind of effective agitprop I could make with AI. Maybe knowing this explicitly will inspire some people to think more about how they can use it too, beyond the hated stereotypes, with thoughtfulness of design put into it. I am still determined to aid in strategic thought and practice, no matter how much mud is slung. Liberation demands it of me.
You’re right to be mad about a lot of things. Help me help you be mad at the right people in a way that will change the world for the better.
Look, people hate AI. No matter how you spin it, they hate it, and it won’t stop anytime soon, no matter what you use it for. I don’t particularly personally care if you use AI to make a couple stick figures but I can tell you that the average worker does. I haven’t met a single proletarian yet who hasn’t despised AI and AI companies. Everyone I’ve met who’s had positive views on it has been labor aristocracy. This is for a very good reason. LLMs are actively being used to ruin proletarians’ lives by causing their wages to decrease, by building datacenters near them and affecting their health and well-being, in many cases even by affecting their livelihood in some other way through collateral damage, and in any number of other ways.
Of course, the post being AI isn’t preventing it from resonating if people don’t know it’s AI, but you’re missing the point entirely. The thing that makes people upset isn’t the output not being perfect, it’s the very fact that you used and supported a tool that’s being used to ruin workers’ lives in order to agitate about something else that ruins workers’ lives. It’s not that it doesn’t resonate, it’s that by using and supporting AI, such as by normalizing the use of AI, you directly betray the working class. That’s why you see such a violently negative response to your use of AI.
I personally know many people who have left this very community for the sole reason that there’s too much use of AI and too many people here are pro-AI. You are actively driving workers away. You are not helping.
Using AI for agitprop defeats the entire purpose. Sure, in this case, most people can’t tell, but if and when the workers you’re trying to reach find out what you’ve made is AI, either because someone can tell (and it is possible to tell) or because someone ran it through a signature detector, they will no longer trust you, and your AI-generated agitprop will at best lose effectiveness and at worst have the opposite effect.
Especially in a case like this with stick figures, I have to ask why? What’s the point of using AI here when stick figures like this don’t take much effort to create in a program like GIMP? Why jeopardize the potential trust workers might have in you in order to save at most a couple hours?
I mean, first of all, I don’t believe that the working class as a whole is going to be put off from commies because of AI use. Your anecdotal evidence does not sync up with mine. In my real life, whenever AI comes up, opinions tend to be mixed at worse and I recall one educator who was even fascinated to learn from me about the basics of how LLMs work.
But regarding the people who are put off, as many people have pointed out rather well over the years in relation to other issues, ceding to the reaction in order to get their trust does not further a communist revolutionary movement. This is fairly obvious to people when we’re talking about a thing like racism, but apparently when it’s AI, it’s harder to grasp. The job of commies is not to go along with whatever is popular or unpopular and cede to its energy and level of education. That’s how you become a reformist who is part of the system in the best case scenario, if not chasing the coattails of fascists. For example, with racism, we aren’t gonna worry because a racist will be put off by us saying no racism. But we’re also going to emphasize they’re blaming the wrong people when they are told to be racist; that them being exploited as a worker is real, that living cost problems are real, but that XYZ group is being scapegoated.
Similarly, our task with AI is to acknowledge, and help with, the real exploitation coming from automation and its “AI” forms, but within this, it will not make us anything closer to a vanguard to simply go along with however people feel about AI and whoever they blame.
This is an argument I’d agree with if I didn’t see AI as nothing more than a mere convenience that not only isn’t required for anything, but actually sometimes makes things more difficult.
Essentially, from my point of view, its benefits are not remotely close to outweighing its downsides.
I’m not saying I’d otherwise use AI but people hate it so I won’t. I’m saying AI is not beneficial to anything overall, and at most it’ll save me a couple hours, so why would I ever use it, especially when it drives people away?
I just feel flabbergasted reading this kind of thing. Do people swear off using GPS too? Youtube because it uses algorithms to recommend videos? What is so different about generative AI that invalidates it to the point of disregarding it as having any possible benefit? Because some silicon valley billionaires swear by it?
I mean, the evidence is undeniable that AI has benefits, including generative AI.
Furthermore, framing it as “mere convenience” vs. “requirement” is an argument that I sincerely doubt would hold up for you, or most people who hate AI, if applied to other things. You would essentially be having to live a monk’s life, where you reject the material in favor of the spiritual, for it to have any consistency rejecting something because it is “mere convenience” instead of required.
That it sometimes makes things more difficult is not without parallel to other types of automation. There are times I’ve used GPS where the timing of the instructions and the delivery of them were so poor that I got more frustrated and confused than I probably would have if I’d just tried to read some directions beforehand and pay attention to road signs. At the same time, it has been enormously useful when I’m in a wholly unfamiliar area and have no idea where to go. If I were to say GPS sucks and throw it out because it doesn’t always go in my favor, that’d be absurd.
That silicon valley hucksters promise the world with AI does not mean that when it fails to deliver the world, it’s total shit. It means they’re full of shit for saying any piece of technology is a panacea. It doesn’t invalidate what the technology actually is capable of underneath that.
AI does have benefits, especially in things like protein folding and the like. I have no problem with it being used for that. LLMs and generative AI have benefits too, just not ones significant enough to outweigh their drawbacks. I do hate how much AI is now used to specifically mean generative AI and I’m sorry for making that mistake myself.
Of course I use GPS, because GPS has drawbacks, but they’re negligible compared to the benefits. Generative AI provides barely any benefit (though not zero) while carrying immense drawbacks that I refuse to dismiss, and those drawbacks are not nearly worth the benefits for me. If your calculation yields a different result, feel free to use it, but don’t complain when you publicize the output and get backlash.
If a use case is found where generative AI genuinely has a lot of benefit and few drawbacks (or at least benefits that outweigh the drawbacks), I would have no problem with it being used for that, but I would continue to have a problem with it being used for other things.
The internet is used every day in service of ruining workers’ lives, yet we’re still on it. This is just the basic “it’s tainted” purist argument that makes no sense in actual practice.
^
I was on lemmy.ml mostly and came here for a change. I was shocked to see the state of AI output acceptance. I struggle to find how it is remotely compatible with Marxist analysis of labor at all. As you say, using it for this sort of stuff defeats the whole purpose in my eyes.