An Incomplete Product Sold as a Tool: How to harvest more data for an eventual Product to strip people of even more of their money and autonomy
Lots if not most people hate AI. There’s no doubt about that in my mind. Ambivalence arises from people who couldn’t care less about anything in the moment that isn’t in their routine of barely making ends meet. You tell them that AI is bad and they’ll ask what about this topic is going to help me finish my two jobs at the end of the week and have enough money to eat something more substantial than instant noodles? If GPT, Grok, Cortana’s frankenstein zombie (CoPilot), and other “AI” is what we’re being sold on now, then what is it going to turn into? This has probably been stated or brought up before but I’ve been mulling over this thought for a few months and now it will not leave me alone so, to page I will put it. TLDR: All of the data that’s stolen and that people are freely giving away by interacting with these current “AI” can easily be used to build general psychological profiles for types of users.
Before I get into this complete thought I have to lay the asphalt. First, allow me a short tangent into the realm of semantics and marketing. I typed “AI” in quote because what we have now, as members of the general populace, is not AI. LLM’s (large language models) are Parrots, as are Generative AI (Gen AI). They will talk to you in the same manner that you talk to them but with the perfect customer service rep’s flair. They won’t talk down to you, they won’t call you out for completely idiotic trains of thought. They are a black mirror. A shiny obsidian-like surface within which we see ourselves but slightly different. A darkness surrounding our visage as we peer into it. They reflect ourselves with distortions. Black Mirror is not just a series to watch, which also gets into the dark recesses and the cross-section of technology and the incompatibility between new tech or technology that’s been perverted has with our psyches. A black mirror is an old, dare I say ancient concept(1) tied to magic and rituals around confronting one’s own mind with purpose and intent.
In scifi for most of history, AI was written as something with actual intelligence. Technology that would converse with you and ask actual questions of you. Synthetic entities that have the capacity to reason, think, and understand things on the level of an adult human but ultimately lacked that spark that separates a very clever automaton and a real living person. A concept integral to properties like Ghost in the Shell(2) and Blade Runner(3). Marketing has taken the idea of AI, and used that to sell us a technology so incomplete they could only advance its growth by crowd-sourcing it’s continued training.
People are wittingly and unwittingly contributing to the growth of something predicated on a lie because people think they can create Frankenstein's monster and control it. People who actually believe in and deeply fear fictional concepts and cognito hazards like Roko’s Basilisk(4), the allied master computer AM from the short story “I Have No Mouth and I Must Scream”(5) , or even Hal 9000 from 2001 A Space Odyssey, or David from Prometheus and Alien: Covenant. Idea’s that ultimately AI will evolve into some malevolent entity that will find pleasure in toying with humans like a child would toy with insects, or an eldritch elder entity with designs beyond human comprehension. People who deeply believe in Accelerationism(6) taken to the edge-lord extreme and Dark Enlightenment(7). People who secretly or not to secretly wish to be full blown authoritarians, each forcing to the world into their distorted and idealized version of society, are whispering in the ears of tech CEO’s and world leaders. You know some of their names, Peter Thiel, Curtis Yarvin, Elon Musk, Sam Altman etc. The tech broligarchs as they’ve become known to some. They fall into a small group of people who believed in their own bs just like someone else connected to scifi, L. Ron Hubbard. The name rings a bell because he is who we have to blame for Scientology. Before that, he was a writer whose stories seemed to give way to an inability to separate fact from fiction.(8)
A well spoken Zealot(9) who believes in their own bs can easily become the best marketing agent any movement can ask for. Zealotry requires that a person’s mind, their entire psychological underpinnings, are reformatted towards, and have assimilated their specific topic to the point that every word they utter is in service to that belief. When you take an idea such as a Cognito Hazard, mix it with the psychology of an L. Ron Hubbard, the professional marketing level advocacy of a modern day zealot who has grown up with influencer and grind culture, you get the actual inception at play. A mind virus that society as a whole is trying to fight off; a mind virus which purports that Authoritarianism is the only way to usher in the next epoch of human advancement. The “woke mind virus” is a myth btw.
With that I believe I’ve covered the basic ideas and will now get to the main point of this tirade.
Google started as a search bar for the internet.
Current “AI” were born from something called machine learning(10). I’m not entirely up to date on the programming aspects but what I do know is that university students, materials engineers, and others in that form of scientific study and research have been working with the idea that if they poured their research into their machine learning algorithms, plus some extra math based on their field, the output would give them something tangible to work from. A product of specific input that would tell them, “Hey, based on your previous work, here is a possibility you could reach.” (11) That as a concept was mind blowing. In theory and in practice it would allow people to work through dozens or even thousands of iterations in order to get to the next best thing that their own data was pointing to in minutes hours or days instead of years. A key phrase there is “Their OWN Data”. They put in the work, the blood, sweat and tears in their research and then put that data to work for them. The machine learning they’re working with has actual value and started with actual value while also being geared towards a specific purpose. There is also data from research that has been put out for free access and utilization by various other people who wish to build upon ideas or apply them in new ways. An easy way to refer to this work is Open Source, though there are many different types of data sets that can be used yet may differ from the definition of Open Source which can get complicated but I digress.
LLM’s on the other hand, from what I can tell and what has been reported on, were borne from mountains, Terabytes of stolen data.(12) They are a highly sophisticated mirror with the intelligence of a preschooler learning how to speak that stunt psychological health and growth to the point of psychosis in some people(13). It’s almost comedy at this point that even this concept was written about in scifi. One such being Cyberpsychosis(13a) a term from the Cyberpunk Universe Created by Mike Pondsmith. The two have their differences as in Cyberpunk it’s the overuse and reliance on Cyberwear (replacement of body parts and systems with technological proxies because of health or other reasons) by a person to the point that they experience a break with reality and suffer homicidal, self harm or other violent mindscape’s they are unable to break free of on their own. Ai Psychosis is when a chat bot feeds into someone delusional thinking or maybe even just genuine curiosity and down a spiraling rabbit hole they both go. A self reinforcing torrent of semi-truthful gobbledygook that can only help some people get an A on an essay as a freshman in high school because the teacher has to read 45 others in the course of one night and they’re all 1000 words.
Given LLM’s and similar technologies specialty, which is hoovering up any and all data it can, there’s something else that I’ve rarely seen talked about though there are some articles that broach the subject. Any algorithm that is designed to sift through and catalog information will at some point have enough data to spit out coherent insights on the data it’s gathered. A chat-bot that converses with millions if not billions of people will, over time, be able to produce psychological profiles of the types of people it has interacted with. The maintainers of these things will of course immediately yell and scream that they are not making psychological profiles of each specific user, which is fair enough. But that only means that they cannot attach peoples names to these profiles. Still, as in regular psychology where there are generally accepted personality archetypes(14) which can be applied to every member of the human population, some if not most requiring meshing of multiple archetypes. To be clear, I’m not stating that the more contested attributions to this idea are fact.
There are enough people who know about the Myers-Briggs and Jungian breakdowns of personality who will vehemently argue that they are shallow and underdeveloped(15) codices put together by people who only half understood what they were talking about; for brevity we all know people exist who are: hard headed, easy going, think too much, want to save the world, and just seem to hate everyone. There are underlying pathologies to discuss and bring to the foreground obviously. My point here is, even as shallow descriptors of segments of society, these are very useful when it comes to targeting people for things like marketing products and services or singling out people who could be categorized as provocateurs, or people dangerous to a specific political movement or government mandate. We’ve seen the tech in action because of the likes of Palantir(16). Sure we can say that in that instance he was talking about killing terrorists. This tech goes so much further. This topic has also been covered in scifi. A little story called Minority Report where people capable of seeing possible future events are used to create arrest warrants for people in the present who are likely to commit a crime in the future. From the wiki:
“Each of the three precogs generates its own report or prediction. The reports of all
the precogs are analyzed by a computer and, if these reports differ from one another,
the computer identifies the two reports with the greatest overlap and produces a
‘majority report’, taking this as the accurate prediction of the future. But the existence
of majority reports implies the existence of a ‘minority report’. In the story, Precrime
Police Commissioner John A. Anderton believes that the prediction that he will commit
a murder has been generated as a majority report. He sets out to find the minority report,
which would give him an alternate future.
However, as Anderton finds out, sometimes all three reports differ quite significantly,
and there may be no majority report, even though two reports may have had enough in
common for the computer to link them as such. In the storyline, all of the reports about
Anderton differ because they predict events occurring sequentially, and thus each is a
minority report. Anderton's situation is explained as unique, because he, as Police
Commissioner, received notice of the precogs' predictions, allowing him to change his
mind and invalidate earlier precog predictions”
You probably know what I’m gong to bring up next, this type of thing is actively being worked on and in essentially beta status in the wild(18). Another instance of scifi warning about a thing with examples but people doing it anyway. I mean it’s just science fiction right? It can’t possibly happen exactly like it does in the book… What’s an acceptable error rate for you? Some future ai robot cop shoots your kid dead because they were literally in the wrong place at the wrong time but “the data suggested that they were there for no good so tough luck, shoulda raised your kid better?” Crimes can happen across the street from police stations, what if then the AI says the cops are implicated? Are a few negative outcomes worth the results? Why does that sounds familiar?
The way we type, the words we use and how we use them can all be turned into fine data points. These data points can then be used even before shoving some psychology textbooks into a chat bot with some tweaks by someone who knows psychology and boom. That bot can then spit out profiles and sub profiles for personality types. These can be sold to companies who want demographic data to sell you more trash that you don’t need. Each ad getting more and more invasive because you happened to tap on one while trying to scroll, or maybe one did seem interesting because that water bottle looked nice, or you were hungry and an ad popped up so you tapped on it to get the 10% off. The data mined from you interacting with ads can only make the data psychological data collated from chat bots more effective.
They’ll predict when you’ll have a breakup. They’ll predict when you have to go to the bathroom. They’ll know how long you spend on the toilet and how long you spend just playing a game to kill time instead of working cuz it’s a slow day and everything is actually accounted for. They’ll know which political leanings you have even if you stay away from politics discourse. They’ll know when you’re about to lose your job and then try to give you ads for loans. They’ll know.
Or maybe they think they’ll know. You constantly scoff and laugh at the absurd ads you get because they don’t know you at all. You’ve made sure to poison your data so the bots don’t know the real you. That just means you’re a different type of person. One who knows how to subvert systems maybe. That’s someone to watch too. Maybe they glean that your low-tech lifestyle which is perfect for you is a problem for them? If you have nothing to hide why don’t you just let them market everything at you unceasingly? You must be hiding something. As I read this it sounds like the ravings of a mad-person but if the year 2025 has taught me one thing, the truth is usually worse.
References: All references provided for the purpose of, if you wish, look into these topics further.
Shalaby, A. Classification for the digital and cognitive AI hazards: urgent call to establish automated safe standard for protecting young human minds. DESD 2, 17 (2024). https://doi.org/10.1007/s44265-024-00042-5
(1) Black Mirror: https://www.ian-leslie.com/p/the-real-black-mirror
https://www.cambridge.org/core/services/aop-cambridge-core/content/view/S0003598X21001320
https://en.wikipedia.org/wiki/Mirrors_in_Mesoamerican_culture
(2) Ghost in the Shell: https://ghostintheshell.fandom.com/wiki/Ghost_in_the_Shell_Wiki
(3) Blade Runner: https://bladerunner.fandom.com/wiki/Main_Page
(4) Roko’s Basilisk: https://www.iflscience.com/rokos-basilisk-the-banned-thought-experiment-you-might-regret-reading-about-78478
(5) I Have No Mouth and I Must Scream: https://en.wikipedia.org/wiki/I_Have_No_Mouth,_and_I_Must_Scream
Book Store/Comic Shop locator websites to find local bookstores and comic shops to support:
https://www.comicstorelist.com/
https://comicbookstores.co/
(6) Accelerationism: https://en.wikipedia.org/wiki/Accelerationism
(7) Dark Enlightenment: https://www.britannica.com/topic/Dark-Enlightenment
(8) L. Ron Hubbard: https://en.wikipedia.org/wiki/L._Ron_Hubbard_bibliography
(9) Zealot’s current definition: https://www.merriam-webster.com/dictionary/zealot
(10) Machine Learning as defined by IBM: https://www.ibm.com/think/topics/machine-learning
(11) Machine learning in Material Science articles from Nature. https://www.nature.com/collections/egijhgcdcd
(12) Gen AI trained on stolen data: https://www.dglaw.com/court-rules-ai-training-on-copyrighted-works-is-not-fair-use-what-it-means-for-generative-ai/
(13) AI Psychosis: https://www.psychologytoday.com/us/blog/urban-survival/202507/the-emerging-problem-of-ai-psychosis
https://www.pbs.org/newshour/show/what-to-know-about-ai-psychosis-and-the-effect-of-ai-chatbots-on-mental-health
https://psychiatryonline.org/doi/full/10.1176/appi.pn.2025.10.10.5
(13a) Cyberpsychosis: https://cyberpunk-red-2047.fandom.com/wiki/Cyberpsychosis
(14) Psychological Archetypes basic understanding: https://glossary.psywellpath.com/types-of-archetypes-in-psychology#types-of-archetypes-in-psychology
(15) Personality Tests Like Myers-Briggs Can Mislead More Than Reveal https://neurosciencenews.com/personality-tests-unreliable-29609/
(16) Palantir CEO Alex Karp says yes, its tech is used to kill people: https://www.inverse.com/input/tech/palantir-ceo-alex-karp-says-yes-their-tech-is-used-to-kill-people
(17) Minority Report: https://en.wikipedia.org/wiki/The_Minority_Report
(18) Crime Prediction:
Iqbal, N., Hassan, A. & Waheed, T. AI-driven crime prediction: a systematic literature review. J Comput Soc Sc 8, 53 (2025). https://doi.org/10.1007/s42001-025-00373-z
Algorithm predicts crime a week in advance, but reveals bias in police response. By Matt Wood https://biologicalsciences.uchicago.edu/news/algorithm-predicts-crime-police-bias
We’re All So F’d | NVIDIA x Palantir Global Surveillance, “Pre-Crime” Arrests, & AI https://www.youtube.com/watch?v=5lYsO4k7OIY
For fun links:
Overview In recent years, with the wide application of open-source LLMs such as DeepSeek and Ollama, global enterprises are accelerating the
The Flock Camera Leak is Like Netflix for Stalkers: https://www.youtube.com/watch?v=vU1-uiUlHTo
Texas attorney general sues major TV manufacturers for being spyware https://www.youtube.com/watch?v=viKdo5T7gZw
