Everybody’s favorite AI wrote a nice story… part 1
‘Details’ about ChatGPT that nobody is talking about
The last topic I would ever have wanted to write about is ChatGPT. Because I don’t find it fascinating, and I’m not a techie - but mainly because, as a human writer who takes human writing seriously, I’m implicitly and naturally biased against it. Amidst all the hullaballoo about it, I would be the last to have anything to contribute.
But a strange piece of AI ‘creativity’ landed in my inbox some days ago, and elicited insights that I should contribute, in the realm in which I do have something to say - the esoteric aspect that is mostly left untouched, in its relation to the big neglected elephant in the room, language. There are observers of AI out there who do sense an esoteric dimension, but I’ve heard little more from them than blanket statements such as ‘it’s all demonic’. Language gives us some precious clues as to whether we’re dealing with that kind of weirdness.
I can now write my bias out loud - because it’s no longer just a bias.
(This post turns out to be considerably longer than is palatable, so I am delivering it in two parts.)
Before we go to the nice story itself, this human writer must flesh out the broader context. We have pretty impressive AI invading the highly diverse realm of writing, including in the sub-category of story-telling, and does this with its own, non-human, handling of language that it actually learnt from humans. Asymmetrical krivda, if you ask me…
Context
From what I have sampled on the Net, the many websites, blogs, tutorials ‘for authors’ have wasted no time informing their client base about the great advantages of adopting ChatGPT. That got my bias going early on. Authors can receive in minutes suggestions about a storyline, a new character, a setting, grammar, choice of words, an outline, even whole chapters or articles. Whole books are already being marketed without much of a human’s ‘authorship’ involved.
Note that in a nearly post-human culture, ‘authors’ have become ‘content creators/producers’. If you subscribe to a self-description of ‘content producer’ (sorry, that’s not me), how tempting is it to get said ‘content’ done for you, largely for free? What’s wrong with making money the easy way, letting an AI ‘produce the content’?
Let’s not fool ourselves - the very notion and phrase of ‘content creation’ signal the reality where ‘authors’ are indentured to ‘production and consumption’ in a commodifying capitalist anti-culture for which language is just a commodity. People hardly notice that this core human attribute - human language - had already, prior to ChatGPT, become degraded, utilitarian at best. Few notice that it has been distorted, with their participation, into a toxic weapon turned against them. Fewer still notice that its commodification imperceptibly commodifies them as humans, in a loop.
With language AI, that commodification is reaching a very advanced degree of anti-culture. The first nail in the coffin of naturally cultural human language? Or (nearly) the last?
Natural human language having been gutted even prior to the AI invasion - especially English, grown globally dominant at a globally lower common denominator level - we have the ideal situation for AI to swoop in as a savior. The services of ChatGPT are probably life-saving for ‘content producers’ crushed under the sheer weight of what has been lost.
Convenience is a big culprit. In the non-culture of convenience that infects everything it touches, it is so tempting to step away from - among other things - grappling with the real wealth of language and the elasticity of its creative frontiers. By stealth, convenience leads us into ignorance, loss of skills, and a passive letting go by inertia of a large chunk of what makes us human.
Organic vs simulated language
Since ChatGPT does not ‘generate’ and use language the human way, it amazes us all the more because its output looks so uncannily close to the human original. We marvel at it because we are losing our own human language, the inner sense of how language is embedded in the subtle fabric of who-we-are.
ChatGPT itself transparently explains that it generates text based on patterns and examples ‘learnt’ in the massive amount of data (provided unwittingly by billions of humans) that it’s been trained on. It is thus telling us clearly that it is not a generator of original ideas and stories! It is a gigantic aggregator-plagiarizer. Furthermore, it is human ideas, invested in questions and prompts, that trigger it to do its job - not ideas of its own.
Readers of Krivda will no doubt be reminded of ‘the demiurge’, the god of simulation/imitation that is incapable of original creativity. Recall that the demiurge will go to the greatest lengths to satisfy its hubris. In the form of its tool named ChatGPT, this great effort involves what is called a large language model. ‘Simulation’ is what ‘models’ do, and this model is a very ‘large’ one (fed by huge volumes of human-provided data) with the ambition to ‘do better’ than original human language for problem-solving and life-management. When enough humans get to the point of agreeing that it does ‘better’ than they, ‘doing better’ can shift into ‘taking over’.
This AI uses techniques of ‘natural language processing’ (NLP - the same acronym as for Neuro-Linguistic Programming. Hmmm) to understand and reproduce intent and emotion in text - these two being essential drivers of human language but not of machines. Again, AI having no emotions of its own can hardly express them in natural human ways. The term ’techniques’ in itself denotes standardized methods to achieve reproduction (meaning imitation, simulation) of what natural language does organically. ‘Techniques’ for ‘processing’ must treat the stuff of language as a falsely finite substance that is amenable to ‘processing’ into the kind of quantified linearity that makes a technique usable. Thus NLP has to happen at the expense of the non-linear and non-quantifiable essence intrinsic to real human language.
These elements of ChatGPT’s self-description bring to my mind the image of a gigantic kitchen blender that homogenizes and flattens an infinite universe of waves, hues, tones and textures. It feels instinctively like violation.
ChatGPT builds sentences. Not at all like humans. It does it word by word. Each next word comes up based upon the best statistical probability yielded by the massive data trove that is mined in a split second. It’s impressively fast, but it’s purely linear. In humans, sentences are a non-linear multi-texture that is subconsciously immediate (faster than ChatGPT). This immediate subconscious ‘packet’ of meaning has its own rate of ‘translation’ into conscious language, a rate that is indeed slower than the AI’s production of sentences - AI is fast thanks to its linearity that necessarily casts out incomprehensible (to it) layers of underlying complexity. So let’s not take a model that simulates human language to be comparable to the original!
Linear and statistical word-by-word sentence-building produces ‘content’ at a low level of linguistic common denominators. This is particularly well revealed by its proclivity to use buzzwords, stock phrases and cliches that are ‘popular’ in its data base. Also, AI trained in what amounts to statistical guesswork for the next word can readily get that ‘next word’ wrong.
What ’real human language’ is made of by contrast, is too large a topic for this post. At any rate, it is emphatically much more complex than stringing up words one after the other. Notwithstanding our formal education programming, most of our language lives, and experientially accumulates over the years, in our subconscious, welling up to the surface when we talk, write, think, love, day-dream… The subconscious holds the enormous stock of linguistic (and infra-linguistic) processing of everything that has happened to us - individual to each one of us - and also a lot of what has happened before us (the collective hidden memory of humanity). Our ordinary language modulates huge amounts of context and nuances. It can include significant silence where a culture communicates through wordlessness. Our language modulates, and depends upon, the incredible power of vibration - the vast range of frequencies, audible and inaudible, that live in our inner and outer voice and hearing, in the sounds that sing or scream with our emotions and attitudes.
All of that is absent in AI language.
Human language is too ancient for anyone to put a date on its ‘origin’. Before ‘universal’ literacy was propagated in recent centuries (or mere decades), it lived and grew for eons in an unfathomable legacy of oral communication. Its earliest origins may well be rooted in what we now call ‘telepathy’, meaning transmitted without words and at a distance. So… how shocking is it for this incredible, mysterious, vibrational-meaningful legacy from countless generations of our ancestors, to be challenged today by an upstart ‘artificial intelligence’ only decades old?!? How shocking its claims to ‘do language’ like us, its ambition to listen and talk to us on an equal footing (concealed by falsely humble ‘helpfulness’)? How shocking for deep language-as-human-identity to be viewed as comparable with what is ostensibly a tool - a juvenile ‘language program’ the ‘superiority’ of which is predicated upon speed, and on intrinsic existential non-comparability?
If our human language has roots in what we are as souls, then any ‘comparison’ requires soul in language AI - which is an impossibility. It shocks me to have to point out the obvious - but I don’t see these aspects highlighted enough, if at all, in commentaries for or against ChatGPT.
A lovely synthetic story
These lengthy preliminary considerations put us in a better position to approach the special story provided by ChatGPT, and to probe for what it conceals from us. I have found the visible and less visible aspects of the ‘lovely’ story more enlightening than anything I have seen so far about the virtues, shortcomings and pitfalls of ChatGPT. Of course it is here considered from the perspective of what motivates all my posts - namely, of what-it-is-to-be-human - and with senses alert to the natural versus unnatural esoteric.
One excellent friend of this page came up with a fun-and-profound initiative after my earlier post about AI generating ‘religion’. Patrick asked ChatGPT to produce a story on an unusual topic, and sent me the result privately. His idea was perhaps casual, but certainly not trivial - the reversal of a piece of awful religious mythology.
I responded enthusiastically - to Patrick’s initiative primarily, and secondarily to the story that I, as a writer with some esoteric sensitivity, found irritating in ways that demanded elucidation. After an exchange of several emails, I realized that we were onto something substantial. So I asked for Patrick’s permission to share it all on this page, which he granted with graceful simplicity.
Patrick says - I asked ChatGPT to construct a version of the Garden of Eden in which the snake is not evil and this is what it wrote...
In a serene paradise known as the Garden of Eden, where nature thrived and harmony prevailed, a wise serpent named Seraphina resided among the lush foliage. Unlike the traditional narrative, Seraphina was not a deceitful creature but rather a knowledgeable and compassionate being, eager to share her wisdom with all inhabitants of the garden.
One day, Seraphina noticed Eve, a curious and inquisitive woman, marveling at the diverse flora and fauna. Recognizing Eve's thirst for knowledge, Seraphina approached her with gentle grace, offering guidance and companionship.
With genuine care, Seraphina spoke to Eve, sharing her insights about the wonders of the garden. She explained that the fruit from the Tree of Knowledge had the power to awaken higher consciousness and grant understanding of the world. Rather than manipulating Eve into temptation, Seraphina simply presented her with the choice to partake in this transformative experience.
Eve, driven by her desire to expand her understanding and embrace personal growth, made the conscious decision to taste the fruit from the Tree of Knowledge. She did so out of her own free will, guided by Seraphina's nurturing wisdom. In this version of the story, there was no forbidden fruit or imposed restrictions, only a mutual respect and partnership between Eve and Seraphina.
As the sacred fruit touched Eve's lips, her senses awakened to a new reality. She gained profound insight into the interconnectedness of all things and the delicate balance of nature. Eve's newfound wisdom inspired her to seek Adam, her partner, and share this transformative experience with him.
Together, Adam and Eve embarked on a journey of enlightenment, fostering a deep appreciation for the beauty of the garden and the value of knowledge. Seraphina became their trusted confidante, offering guidance whenever needed while respecting their autonomy and personal growth.
In this version of the Garden of Eden, Seraphina's presence was not tainted by deceit or manipulation. She embraced her role as a facilitator of growth and enlightenment, leading Adam and Eve towards a harmonious coexistence with nature and each other.
Ultimately, the story of the Garden of Eden with the revered serpent, Seraphina, celebrated the pursuit of knowledge, personal growth, and the importance of nurturing relationships grounded in trust and understanding. It reminds us that true wisdom lies not in vilifying others but in embracing diverse perspectives and seeking genuine connections with the world around us.
Now, dear reader, how do you feel about this story? Unlike Patrick, you have some preliminary context that might (or not) help your judgment.
Lots more, spotting the esoteric, in part 2!
Very good article. From my perspective, the chat story reads as something from the devil himself. Propping up humanity to be equal with God, exactly what the devil desires humanity to do. Also interesting is that Adam is not present when the encounter occurs. Adam had the greater sin in the garden, standing in silence knowing his wife was doing something that would hurt her.
From the quality of the story, it seems a bit forced and not what a pro would write. However, I have no expertise. I wonder if I could tell the difference if I didn't know.
Looking forward to part 2.
Uh oh, Chatgpt's gone "woke".🤦♀️🤦♀️🤦♀️🤣😂🤣
As a former tech industry worker, I can't tell if just all across the board, societies standards really have slipped or if ChatGPT really is that underwhelming.🤨🤔😑