JR's Zampanio Branch Ai Policy

TLDR:
  • the vast majority of my Zampanio fanworks have no LLM and I try to clearly mark where it does
  • None of my other fandom work has LLM at all (Homestuck etc)
  • i definitely get LLM have gone from fantasy horror to uncomfortably real horror, which is why I am clearly labeling it and making it harder to find
  • its okay to leave. Zampanio needs you to live a long, healthy life.

    • Hi yes hello, it's me, JR, your friendly neighborhood spiral faced horror/waste.

    I wanted to be clear about how I use generative AI, and more importantly, how I don't. (and unfortunately for us all when I decide I want to be clear I get hella wordy, lol)

    I find LLM (large language models, another word for generative AI) to be unsettling, or even dangerous! "AI Psychosis" scares the shit out of me!

    I also find the WAY LLMs are currently being used (to replace human workers, to make even more profit for billionaires, to make it easier and easier for people to stop talking to each other and learn new skills) high key terrifying as well!

    So I do what I always do with my fears: I make art about it. Art I try to keep as ethical as possible, with as much transparency as possible. I make horror content about obession, about losing yourself to something so utterly you can't even recognize your past self. I'm no stranger to the legitimate dangers this can cause just as a topic to explore. I try to have content warnings, to give people guardrails, to learn about dark patterns so I can specifical disrupt them. So this rambling 'policy' is kind of me trying to do the same thing for AI use.

    My goal with having a small amount of LLM content in my branch of Zampanio is to play with the ideas it causes in me. IS there a safe way to engage with something like this? The early LLM models, like Deep Dream, were so fascinatingly liminal, like half remembered dreams. Can I, as a human, make content that invokes that, using LLM images as a tone setter? I want to use my art as a way to familiarize myself with something I find scary. What can LLM actually do, and how worried do I have to be about its place in my future?

    If I use LLM generated content (or take LLM content donated from a fan) I try to make it as clearly marked as possible. Either old type really bad and uncanny images or literally screenshot it on the chatgpt page or whatever. They tend to be either on eyedolgames.com or if you get deep enough into the bathroom that it starts breaking down (if you know you know).

    I code and write myself. I make dioramas these days, also by myself. No AI touchups, often even no spellchecker (I embrace the typos, lol). I collaborate with writers, artists and musicians who, to my knowledge, arnen't using LLM. I also collaborate with specific LLM enthusiest I know are using LLM as ethically as anyone can these days.

    I add LLM content deliberately in places where I feel it works, not as some kind of normal part of my creation process. Usually weird obviously AI images in infinite fake websites. Once I trained an AI on my own voice and put that in the bathroom as "Truth"... Past me making hundreds of audio logs... yeah. My voice is gonna be stolen by a scammer. I wanted to explore what that would feel like before it really happened, you know?

    I'm not aware of any ai generated music in my branch of Zampanio. I think theres an LLM podcast about Zampanio somewhere? Not anywhere you find in a normal playthrough, somewhere I just hid for myself to remember how it felt to hear these manic people talking and hallucinating about something I'm deeply familiar with... I wanted to undrestand how to detect it when it WASN'T something I knew about? If that makes any sense.

    Any AI generated images are intended to be as obviously bad and fake as possible. I know actual artists irl, they're not gonna let me get away with (if i even wanted to) passing off ai slop as their work, you know?

    I'm not aware of any ai generated writing in my branch of zampanio, barring times where I am deliberately asking an AI something about zampanio and take screenshots of its response or otherwise label it as an interview with an AI. (all those are in eyedolgames.com, pretty sure.)

    I use local models running on my own computer and not some datacenter (as does my main human contributors who use LLM) rather than stupid corperate ones that are bad for just about everyone. I specifically use local LLMs to generate uncanny, unsettling, images that I feel like there's no real analogue from human created? The point of eyedolgames.com is that its run by inhuman monsters absolutely sure they are aboslutely nailing blending in as human. They create eerie dream like infinite sites they think appeals to humans, of which they absolutely are.

    You can see examples of the types of images these local models makes, here.

    Work has me using Claude, specifically to evaluate it for company wide use and I find it to be... only useful in very specific ways that don't even need a lot of computing?

    Mostly I just had Claude play zampanio or read source code and summarize it for me. I find it fascinating the way this alien mind, pruned into a customer service shape, engages with it? It let me do my job (evaluate Claude for our company) while also creating art to process how I was feeling about this becoming my job.

    You can see examples of that, here.

    My branch of Zampanio is a maze. Not all branches of Zampanio are, but mine...really really really is. That means that not even the current version of me knows where everything is, and what does or doesn't exist. I might be missing LLM content, and if you find any you want me to more clearly mark, feel free to let me know and I might fix that.

    I can't promise I will. I can't promise I'll still even be ABLE to. The rot takes all in the end and there's going to be things I"ve made that I can no longer edit.

    It's okay if the way I'm engaging with AI doesn't work for you. It's okay to feel uncomfortable and simply leave. I'm making what I'm making because it feeds me. I hope it makes the people who find it happier, more confident about learning new skills, and healthier. That's what it does for me. If it's not doing that for you? Thanks for caring enough to read this far. Obsession is a dangerous thing, its always the right move to stop playing.




    Some thoughts after sleeping:

    eyedolgames.com isn't JUST about my fear of LLM.

    I was never really using llm because i had no other options, you know?

    i bought stock photos for Zampanio North, for example, for the CCTV stuff. (if you know you know)

    Instead its to make art about enshitification. I made Zampanini because I got spooked by doordash showing me weird food, and its only gotten worse, for example.

    I made Eyedlr when the porn bots were threatening to take over Tumblr.

    I made jackElope because of how shit dating sites have gotten.

    I made the Quotidian Adoption site when looking for kittens myself and having to sift through all the scams and fake pets.

    My goal wasnt just "creepy", though yeah, obviously that, too. It was to create something that captured how I felt to increasingly not be able to trust the Internet, when it had been my refuge.

    For a website to look me into the eyes and say "You like this, right? You'll pay more money because we are doing this to you?".

    To create art about that being the point, about it being an interconnected alien mind performing "human" at me to try to please me... Without having any understanding of why that is hurting people...