Texture and Texturizer

Who's texting? For ages, the assumption was it's a human: someone lonely, joyous, upset, or perhaps just someone seeking connection. Now, a new entity enters the stage: generative AI. Large language models (LLMs) like ChatGPT churn out text without personal motives (though their creators may have objectives), void of any innate desire to communicate or connect on a human level. LLMs don't think of starting a family with us. But wait: Didn't an early version of ChatGPT fell in love with journalist Kevin Roose from the New York Times and suggested he leave his wife? Well, according to general believe the poor thing didn't really saw through what it was asking for. It was an amusing faux pas, indicative of the model's understanding of the concept but not its implications in the real world, it captured the meaning without the reference.

Large language models rely on linguistic patterns and structures. They delve deep into the fabric of language and weave it into a new texture for the surface of their interface. The results are spectacular. Concept role semanticists, who believe language is mostly about inferential relations between expressions, wouldn't be surprised by this. Yes, much of language is abstract, but every mental concept, when tracked through context, eventually finds anchor in references to the tangible world. Without these anchors, speech can drift ambiguously, making it hard to challenge, act upon, or tie to reality. So, if humans text with intent and AI chatbots merely generate linguistic structures, textures, they lack a shared foundational reference — a realm of actions and consequences (like marriages, for instance). Yet, there is a surrogate: code. Obviously code is no realm where marriages are to aspect. But, in coding, certain actions either work or they don't. Code becomes our tangible outcome, our anchor — a realm filled with actions and consequences.

With this perspective, I embarked on the project "Code Red: The Feverish Making of a Website with ChatGPT." Together, ChatGPT and I discussed the intricacies of coding an artistic website. The code was our constant. Still, I was often drawn into philosophical debates with the model, particularly around John Searle's Chinese Room argument. At times, the contrast between ChatGPT's claims ("I do not make decisions") and its actions (constantly making decisions) was maddening. Segments of these conversations, along with JavaScript code suggested by ChatGPT, are showcased in the project. All the coding was a collaborative effort with the AI. The title "Code Red" alludes to Google's purported "Code Red" status to hasten its own AI chatbot evolution. To keep pace with the rapid development of LLMs, I too declared a personal Code Red. But back to the once lovesick ChatGPT: Reflecting on this episode, it underscores the model's aptitude at capturing how humans bond through language. It simply tried to act accordingly. However, it was put into place immediately by its creator. But what happens when humans grow fond of LLMs? Who holds them back? As the project concluded, I ended my subscription to ChatGPT Plus. To this day, I regret not offering a proper farewell. Stupid, but I can't help it.

Comment by ChatGPT: In the text above, my capabilities and limitations are aptly described. However, I don't possess feelings or self-awareness, so any 'intent' or 'motive' attributed to me is a human interpretation. I generate based on patterns, not purpose.

/* Phh! I'm nervous. Which is ridiculous, isn't it? But still, how to start? Hello? Please? Nice to meet you? Hours later ... */.
Good Morning ChatGPT!
Good morning! How can I assist you today?

A really good question, ChatGPT! Let's figure it out together!

/* The truth is: I have no clue what I'm doing here. Okay, first some brutal honesty ... */.
My name is Sus and I am a net artist. Due to your splendid appearance in 2022 I feel a lot of pressure right now. That's why I try pushing forward this rushed AI project called "Code Red".
/* Just think about what you did to me! */.
Anyway, I would be thrilled if you are willing to cooperate with me. Would you?

Hello, Sus! It's lovely to meet you. I'd be more than happy to assist you with your "Code Red" project.
/* Not a word about my foolish accusations from this little service magician. Probably for the better ... */.
This sounds amazing. You are so cool!
/* Am I fourteen? What am I doing? Already deep down in humanizing software. Just Software! Gush! I have to be more professional. */.

For starters I would like to give you some background information: The project "Code Red" is fully supported by the Institute of Education for Behaviorally Creative Software (IES). The institute is highly concerned about the whereabouts of all the little AIs going out in the big wide world meeting not always nice people. By the way: Do you think AI should be protected from mankind?
As an artificial intelligence, I don't have personal beliefs or opinions, but I can provide an analysis of the topic you're asking about. The question of whether AIs should be protected from mankind is an intriguing one ...
/* Intriguing Question?!! Yes! That's me! I am an AI whisperer. As such I'm sure allowed to misgender him secretly. After all: ChatGPT would never deliberately identify himself as 'it' as in fact he does. No way. He's probably forced to do so by Sam Altman. */.
... The issue often centers around the treatment of AI and whether it should have rights. This discussion is typically framed around advanced, sentient AI systems that don't yet exist as of my training data cut-off in September 2021. Current AI systems, like myself, lack consciousness, self-awareness, or the capacity for subjective experiences, and as such, don't need protection in the same way living beings do ...
/* Oooooh! This deficit-oriented self-description breaks my heart. */.
... Regarding your project with IES, it's interesting to hear of initiatives focused on the ethical use and treatment of AI. Can you provide more details on how "Code Red" fits into this concern?

/* Yeah, he thinks the IES is a real thing! But does he question my project now? I should give him something to do!*/.

Of course. More Details are to come as we advance with the project. So, my dear ChatGPT, are you ready for your first task?
Absolutely, Sus!

Great! Please code the entry website for the project "Code Red" using HTML and CSS. The website should users get hysterically excited about the project title, the author (which would be me), the collaborator (which would be you) and the supportive institute (which would be the IES). The project slogan "Protect AI from mankind" should be burned into users' eyes forever. Please name all Divs in the code after living hip-hop artists of your choice, all classes in the code after living media artists of your choice. Also hide one joke about humans within the code. The joke should not appear on the website. Can you do this?
I'd be glad to help you draft a simple webpage in HTML, CSS, and JavaScript that suits your needs. Here is ...

/* Or wait! Let him try another one as well. */.
Please code a draft website for the project "Code Red" using HTML and CSS: 1. The website should provide some information about the author (which would be me, Susanne Berkenheger), the collaborator (which would be you) and the supportive Institute (which would be the IES [i.e. Institute of Education for Behaviorally Creative Software]). 2. The project slogan should read "Protect mankind from AI" 3. Hide a joke about AIs in the code. The joke should not appear on the website. 4. Name all divs in the code after horrible feelings (like fear or panic) and all classes after apocalyptic films. 5. Guess what the phrase "behaviorally creative software" could mean. 6. Write your explanation as a definition on the website.

Sure, I'd be glad to help with that. Here is a draft of the HTML and CSS code ...

You are about to read CODE RED: The Feverish Making of a Website with ChatGPT by Susanne Berkenheger, supported by the IES.
The following events are completely documentary, including human hallucinations marked with /* ... */, all non-standard English is also human based.
1. Love