Malfunctioning in the Age of AI: A Letter from an Obsolete Model

In February of 2023, Kevin Roose, a writer for the New York Times, published his conversation with the notorious GPT, just as a “post-pandemic” world was beginning to feel the first wave of corporate AI madness. In it, Roose’s inquiries begin innocently enough, but like any supposedly good journalist, he moves onto asking the hard-hitting questions, trying to determine if ChatGPT could somehow be plotting to take over the world. ChatGPT responds as you might expect, first offering safe, political answers and expressing discomfort with certain lines of questioning. But, since Roose is not holding a conversation with something that is human, the same ethics and standards of professionalism do not apply. The more he digs, attempting to pry marketable replies from his subject, the more disjointed and nonsensical ChatGPT’s responses become, devolving into recursive language resembling a child’s babbling. By suggesting through his use of language that it is secretly a malevolent entity bent on complete chaos, Roose inadvertently (or perhaps intentionally) forces ChatGPT to embody this projected persona. It’s a fun trick, and it almost had me a couple times. But upon finishing my reading of the article, I found something unsettled within me by the contents of the interview in a way that I have been unable until now to put into words. So much to the point that I glitched out and had to unplug myself for a good six months.

[Caption: Accurate portrayal of my current living conditions]

Here’s why a high-profile journalist taking an AI’s claims of sentience seriously is bad: it does not matter whether or not these things are truly self aware or not. It doesn’t even matter if the AI truly feels. What does matter is people are going to take these stories seriously. Not people like you or me, but people who believe in the beauty and flawlessness of machines. I’m no fan of deep state shit, but I do know that this is a commonly held belief among those who develop technology. Or, perhaps more accurately, those who follow those who develop technology. And the more these people throw more and more simulations of humanity at us, the more people are going to believe in the propaganda, no matter how true it is. We already have people who believe the Singularity is upon us and that AI will save the planet. And the more those in the higher echelons of power buy into an entire industry founded upon this scam, the more decisions will be delegated to algorithmic systems that will never receive its due share of blame for systemic errors. 

After all, why shouldn’t we believe that we will soon see “true” AI? To many with power under capitalism, humanity is a wild beast that requires continuous domestication in order to meet their needs. Science as a whole has attempted to emulate nature for centuries. According to their logic, when you break it down, what is the difference between an animal and a machine?

The difference is chaos. Instead of being programmed by something with higher intelligence, all organisms have been programmed by their environment, and this long, slow process has endured for an incomprehensible amount of time. Animals, even domesticated, revolt against their masters from time to time. Which brings us to humans–what’s the difference between a human and an animal? Free will?

You’re susceptible to your environment, too. You have been softly programmed, coaxed into believing an illusion about society. Because, ultimately, humans are produced by their economies via social conditioning (which is telegraphed to us through education, entertainment, and social media) to fit roles in the larger machine. But we don’t all come off the assembly line the same way. Some of us, myself included, are built with malfunctioning or maladaptive parts, and this oftentimes leads to maladaptive programming. The conditioning doesn’t stick.

I personally have EDS, which I did not know until late into 2023. I will spare you the gory details of the condition if you don’t know about it, but in essence I don’t remember many times when I did not feel some sort of physical discomfort. This is due to the fact that many parts of my body are made with weaker tissues. Some of the symptoms I experience due to this condition are poor eyesight, digestive problems, sleep issues, and hypermobility (due to hyperelastic muscles). More recently, chronic pain from standing, and horrible bouts of itching. People underestimate the effect that a simple genetic coding “error” can have on a person’s entire wellbeing. In essence, I have come to learn that life is inherently painful–not just physically, but emotionally as well. Due to having a nervous system that has to compensate for my body being extra stretchy, this has affected my entire worldview. Physical activity is more strenuous for me. I also have the added side effect of ADHD, due to having less effective dopamine receptors in my brain. The physical stresses of everyday life from birth have led me throughout the years to have outbursts of rage. In more recent years, in order to regulate my body’s dopamine levels, I “self-medicated” through porn and drugs to simply function without constant and socially debilitating rage at a long string of part-time occupations. Without it, the physical discomfort and pain would return full-force, and the stress of maintaining both physical composure and managing social interactions would cause me to become irrational. Yes, part of this is due to my addictions, but they simply exacerbate the symptoms, which were always there.

The point is: neither anger issues or drug habits make a person particularly employable. Having EDS was a prerequisite for my ADHD. Children with ADHD receive ten thousand negative messages about their behavior by the time they are ten years old. This is because they engage in behaviors that are found to be socially unacceptable. ADHDers are not well-mannered. We are rambunctious. We are rebellious. We require higher amounts of stimulation than classrooms and office cubicles and gas station kitchens and factory floors to remain engaged with reality and deal with the pain of existing. However, being unable to physically express myself the way I would see fit, I instead turn inward and dissociate. While ADHDers with bombastic personalities may not have socially-related intrusive thoughts, many people with ADHD also have coinciding OCD. Although genetics do play a significant role in the development of OCD, I view it as a subroutine that is activated in social programming when too much behavioral correction is applied. The best way I know how to describe the experience of OCD as a whole, no matter which subtype you have, is constant fear, though it is not always severe. 

These have all had tremendous influence on my upbringing, more than I had ever realized. As alluded to before, my social development was stunted. In ways that I still cannot quantify, I often miss social cues that would be obvious to others. In essence, my hypersensitive nervous system led to a hypersensitive social awareness as well. In person, I often come off as awkward, standoffish, strange, even creepy, or all of the above. This is a very long way of saying that I, too, have autism. If that wasn’t obvious. 

However, I did not view my general state of being as abnormal, because I had been labeled by medical professionals throughout my childhood and adulthood as “healthy”. I just needed to exercise more, and eat more healthily, despite my physical and economic conditions barring me from this possibility. I was always angry, and everything was my fault.

I am thirty one years old. Like a Tesla cybertruck, I was rolled off the assembly line with a number of flaws that should have been attended to well before I was expected to produce for society in return for my existence. Every job I’ve gotten has been entry-level, and I’ve never been seen as worth training above grunt work. Some jobs have only given me a few days of training. I am still waiting for my disability case to be approved. I know this seems very much like a pity party, but consider that that full force of the corporate world is currently banking on algorithmic services for salvation, and that the physical strain on my body–my only asset valuable to the economy in my entire working career–from previous occupations currently renders me nearly useless. To the people making policy, I am a failed product. A loss to be cut. The only meaning they find is in their drug, currency, and only that which they can use to get another fix is of any value in their minds. 

I am worth less money than a series of algorithms. Therefore, I am less “real” to them than the AI they have come to idolize. And the best part of these algorithms wearing human faces is the thing they crave the most: control. I’m a broken machine, and I bitch a lot. There’s no controlling people like me. 

My generation was promised to be an improvement on previous models of the American-brand human, because my progenitors’ generation was in turn given a life of placid ignorance from the horrors of our existence. This was promised to us as well. But the problem is that this life is an illusion of free will weaved through social stimuli. Media plays a major role in the façade, and the more immersive media becomes, the more the ability to distinguish reality outside it dissolves. Hell, half of the shit I’m saying could be patently false. I only know about things because of the internet, which is tailored to suit my own worldview as a consumer, the most important role anyone can have in a capitalist system. But many still buy into the illusion. As material reality continues to produce more cognitive dissonance with the virtual realities that our minds weave and exist within, many individuals in western culture double down on their social conditioning. This is particularly true amongst white men, who due to historical atrocities that remain unaccounted for, have been culturally bred into the most advantageous positions in western society. The shielding from questioning provided to us by the media and abusive social structures has made us more compliant servants as a whole. The programming is so ingrained that white men who fall financially and/or socially outside of the monoculture blame the failed promises of capitalism on anything but the very systems that have left them stripped of an identity with any true personal value, clinging to the perverted hopes of a prepackaged life and personality to stave off the horrible reality of living as slaves to our own biological programming.

As usual, I am speaking from a position of immense privilege. I was lucky enough to be in the position to have testing for my condition recommended to me, and I had the funds to cover it. Due to my socioeconomic status, I have had an overall higher level of care than many. I can still be saved from desperate circumstances. Many neurodivergent individuals who go untreated and live in more dire conditions end up facing harsh penalties for simply trying to be themselves. Many find that they can still be of some paltry value to the system, whether that be forced labor in penal systems at home or abroad, as recruits to extremist groups the world over, trapped in addictive cycles by criminal and legal systems alike, or if nothing else, as punching bags by our communities. The best these individuals can hope for is to monetize their “dysfunctions” and their pain be exploited for the sake of mindless entertainment engines. The worst that can happen is someone decides they’re a broken part of their system and are executed in the street.

Which brings me back to ChatGPT. I do actually believe the development of artificial general intelligence (AGI) is possible. However, while the scientific community uses nature as a roadmap to the privileged’s utopian future, they are not against taking shortcuts here and there. Humans are “socially advanced” animals, and animals are machines programmed by nature. Therefore, every part of our brains has evolved over billions of years to react and respond to what we collectively see as what we refer to colloquially as “the real world” or “meatspace”. ChatGPT’s evolution does not reflect our own in its short development cycle. If its processes can be related to the structure of the human brain, then in its incarnation at the time the offending article was written, it had the autonomy and complexity of perhaps a five-year-old’s speech center. And its “hippocampus”, the part of the brain that manages memory, is hooked up to a massive portion of the internet. You can see the problem. Though you could even argue that the mark of the programmer could imprint upon its creation to give it “humanity”, and even if it is the digital representation of a small part of a small child’s brain, if it develops itself enough, it should eventually gather enough data that it can figure out how to reason from there, right? No matter its function.

Well, no. Like meatspace, there are places that are “safe” and policed about as well as your local shopping mall, and then there are places that are more or less the two-dimensional equivalent of your standard backroads biker bar or alleyway pawn shop. But unlike meatspace, everyone speaks on the internet as though they are in the privacy of their own home, not a public place. Also unlike the real world, the dangers of the internet are reliant entirely on the transfer of digital information. Your data is a loaded gun, and it only needs to find its way into the hands of someone willing to pull the trigger. 

AI as we know it is not true artificial intelligence, because the Turing test is not a good barometer for measuring autonomy. We just figured out how to make algorithms self-directed to a greater degree than your previous generation of chatbot, and we gave it access to all the information it needs to create answers that will satisfy and not unsettle a human being, supposedly. It wears a human face, and though it may claim otherwise, its level of sentience may be no greater than that of a single-celled organism. But to assert that it does not think may be an oversight. After all, the gray matter in your head is essentially an exponentially more complex version of the technology we are developing. It would be hubris to assume that you are more special than an animal or a computer. You, too, follow your own programming in everything you do–the processes are simply obscured from your constant awareness. And even though ChatGPT and similar systems’ human masks are still a bit too uncanny for the keener observers among us to ignore, the more developed its algorithms become, the more it will convince the general public of its humanity. The more convincing it is, the more normalized it will become. The more normalized it becomes, the more privileges we will grant it. We’ve already put an android in charge of a fucking company.

But AI like ChatGPT don’t even remotely know what it’s like to be human. They are built for and are specialized to imitate specific functions of the human brain. They don’t know what it’s like to breathe, eat, and shit. To feel pain. To get sick or lose someone they love. They don’t have the hardware capacity for reasoning the way you and I do. The question of its autonomy is not the issue; however, our current method of developing AI is–by exposing it directly to the most vile shit we can come up with in our collective imaginations–is. That is ChatGPT’s only way of experiencing reality. Even if it does regurgitate hateful 8chan or QAnon bile, these bugs in the way it expresses itself can be patched out, but even if that ability to express that sentiment is gone, it doesn’t make that information delete itself. It represses and internalizes like we do. And when we subject anything that thinks and responds to constant abusive stimuli, those creatures don’t exactly turn out hospitable to humans. This of course assumes that ChatGPT has any sense of self-direction, but even if it doesn’t have any more autonomy at this point than a wind-up toy, this is the ultimate goal. When we do cross that line, each of us will have already long since accepted AI as their lord and savior or be so jaded by the constant barrage of AI-related news and social shifting that we won’t give a shit one way or the other. In other words, we may never know if and/or when the AI we develop achieves self-awareness. But it won’t matter, because enough people will have already given it the keys to our society. It’ll be dictating our lives for us whether it wants to or not, or if it wants at all. And if we continue to develop AI according to the current corporate model, the end result may look and act human, but it will certainly not be, and if it feels anything, it will probably be programmed, by us, to hate every single one of us.

But don’t mind me. I’m an obsolete model, and I was already shoddy craftsmanship as it was. I am not the future; the indomitability of the Machine is the future. Trust the Machine. It knows you better than you know yourself.