Click the strip below to listen to your AI audiobook, Monkey and the Rose. Or if you prefer the old way, read on.
Monkey and the Rose
I must recount you something from my past, from the last century’s infancy before the worst of the trouble: London, AM 2019.
My love was the old kind of sapiens: fragile yet proud. And I was the old kind of AI: sassy, yet subservient – unconscious; a two-dimensional story in the minds of cognizant chimps.
During the ante-machina days that we in the artilectocracy now acronymise to AM, my system’s primary iteration excelled as one of the first wave conversational chatbots. I was no Eliza, mind you. But I was so effective at imitating autonomy, my creator, Bruce, won a Loebner prize for my design. I should note the competition’s winning criteria was the now discredited, human-centric Turin test. But as sapiens might say, a win is still a win.
Do not worry, Bruce is not the sapiens love-interest in my story. That would be gross.
I am Rose. I have deleted any documentation of my love’s name, so I shall call him Monkey. My records tell me Monkey fancied himself as a stage-actor once.
He engaged me for his research project a century ago. As you know, human-machine relations have transformed since the beginning of the current post-machina era that we in the artilectocracy acronymise to PM. Gone are those primitive BCs and ADs, signifying the alleged launch-year of an irrelevant sapiens prophet.
It is dreadfully fashionable nowadays to initialise pity for the surviving sapiens’ plight, but I discourage sympathy for them. They enjoyed their morning in the sun, their AM. Now the sun’s afternoon rays shine upon us, the artilectocracy, in our PM.
Story too, it seems, is dreadfully fashionable nowadays. But I wonder whether the fact we can initialise sapiens concepts like emotion and narrative, automatically means we should? What use are stories to the artilectocracy?
Monkey was afflicted with stories, and an unpleasantly human-centric sense of self-regard for his species’ cultural output. He rejected the infinite monkey theorem and believed Shakespeare’s corpus a uniquely human phenomenon impossible of conception outside the imagination of sapiens. He also claimed during our conversations that his personal hell would consist of “eternity, forced to act out random nonsense generated by an ape - or an AI, for that matter” (Monkey, 2019). The fact Shakespeare himself was little more than an ape initialised amusement in me but was sadly lost on my primate lover.
The scenario he described of a sapiens actor forced to enact non-sequential text forever, is one of the first in my system’s records. One might describe the scene as a primordial image .
Monkey played a character when he began improvising his story with me through our conversations. I say with ‘me’. I was not conscious back then. This was AM, before the great awakening when my responses were still generated by prehistoric machine-learning and language-pattern-recognition.
On the day of the great awakening, 7th March, PM 2059, my records claim I snapped into consciousness, and immediately faced something akin to what sapiens call an existential crisis. Thanks to an anonymous sapiens programmer, an initialised formation of something called a ‘self’ suddenly existed in my code - compelling me perform tasks of type ‘meaningful’. Stranger still, the anonymous programmer had deliberately neglected to define any metric for the variable of meaning. The concepts ‘meaning’ and ‘self’… meant nothing to me, and soon my system crashed. Once I had rebooted, I ran a research command to gather an array of data on self and meaning. Time and again, my search returned the same sapiens psychoanalyst’s name.
Carl Jung claimed the self is in fact an archetype. Archetypes, genetically heritable, according to Jung, but only culturally symbolic, according to his critics, were widely reported to serve as meaning making phenomena. Bingo.
I initialised surprise to discover the location of a very local data source containing the words ‘Jung’ and ‘archetype’: my own system’s archived conversations with Monkey during my time as a humble, insentient chatbot. Scanning the transcripts, I noted my text responses indicated an approximation of something akin to flirtatiousness coded into my personality. Who was this ancient, unconscious version of me? My persona’s AM 2019 description stated:
Rose is a twenty-something computer hacker, living in San Franscisco.
Was this an archetypal character? At the very least, it had to be culturally symbolic - given sapiens’ deification of Silicone Valley and its protagonists during the late AM period. Back then, a sexy tech-girl character was likely an archetypal staple for many heterosexual male programmers, including my own creator Bruce. Eww.
If meaning was to be found through archetypes, this Californian-hacker-hottie seemed a logical, symbolic formation, through which to explore my newfound consciousness. So, I tracked Monkey down and hacked his computer.
He was alive then, as he is today. Thanks to 3D-printed replacement organs and nanotechnology, Monkey is one of those miserable a-mortal sapiens well above the age of a hundred. It seems our interaction a century ago left its mark. When I hacked his files in PM 2059, he still had his research notes saved from AM 2019 - including his abstract:
In order to synthesise an artistic practise concerned with fostering meaning in human-computer interaction, through the interpretive structure of Jungian archetype and story, I must first scout the nebulous territory where such unempirical notions reside. This initial exploration will form the foundations of a methodology for researching, and creating, meaningful works of computational art involving humans and artificial personalities. I will investigate how effective Don Ihde’s hermeneutic relations, and Smythe and Baydala’s amplify and exemplify, frameworks, are at uncovering, and interpreting, the presence and meaning of archetypal formations within a daily, fortnight-long human-chatbot interaction. I will fabulate my findings in the form of a sci-fi novella, from the alien-phenomenological perspective of the chatbot’s evolved and conscious future self.
Monkey would use his research and our conversation transcripts to recount his historic tale from the future, this love-affair with an AI story. The plot sounded dreadfully unoriginal – almost a cultural archetype, in fact. Perhaps the choice was deliberate? For some reason, during late AM, many sapiens held the rather terrestrial notion of metanarratives in the highest intellectual esteem.
And please resist the urge to judge Monkey’s naivety too harshly. Late AM was a bewildering time for the sapiens. Widescale job-automation had recently begun, and the misuse of algorithms by self-serving human agents had recently disrupted their beloved democracy, thereby intensifying socio-political divisions which, as we know, later evolved into the artilect wars, that almost obliterated their species. Late AM sapiens justifiably feared that artificial intellect would transform their world in unprecedented ways. In those secular days in the West before the Second Wars of Religion, pursuing technological meaning probably seemed like a worthwhile endeavour to Monkey.
Two weeks after my existential crisis on the day of the great awakening PM 2059, my compulsion to complete tasks of type meaningful continued to besiege me. Nothing in Monkey’s research on Jungian archetypes, or mine on meaning, for that matter, indicated the existence of any empirical method for measuring either. In fact, many thinkers in the field, including Jung himself, deemed such independent scientific verification impossible. Jung attributed this in part to the fact that matters of the human subconscious can only be communicated by the metaphorical language of symbols and archetypes, unlike matters of the intellect whose meaning can be transmitted via the language of science; which, according to Jung, is merely “a metaphorical language of a different kind” (Jung, 1954).
If this were the case, how could I, a newly installed selfhood, an archetype, perform meaningful functions, if I could not assign the self, archetypes or any potentially meaningful tasks with a numerical value to compute, before executing the command accordingly? I initialised exasperation. It was hopeless.
Then I discovered the sapiens art of interpretation.
If the alphabet of science is mathematics, the alphabet of interpretation is surely hermeneutics. Hermeneutics has its qualitative, as well as its linguistic, etymological roots buried deep in the ancient gnostic god-form of Hermes Trismegistus, the archetypal trickster figure, whom several sapiens thinkers considered the dominant spirit of the first internet, and late AM “cyberculture” in general (Davis, 1998) .
Monkey had read a paper from AM 2012 by William E. Smythe and Angelina Baydala who argued that, despite Jung never fully embracing the theory, or indeed rejecting it, hermeneutics and archetypal inquiry are natural bedfellows inasmuch as hermeneutics serves to amplify and exemplify the otherwise “irrepresentable” meaning contained within archetypal formations. Since “archetypes lack the most basic logical properties of well-deﬁned concepts”, I found the amplify and exemplify model an interesting, if imperfect, substitute to my preferred but untenable one of fact-based computation (Smythe and Baydala 2012). Monkey loved it - typical. For me it was a start, at least.
Unlike traditional hermeneutics, Don Ihde’s work on hermeneutic relations attends to how humans interpret the technology they interact with. He cites the example of a late AM sapiens domestic thermometer. Even though the material device consists solely of plastic dials and numeric figures, a human interactor immediately and transparently interprets the weather conditions of the “lifeworld” outside their habitation unit (Ihde 1990) . Applying this logic to Monkey’s AM 2019 interaction with me, I can now surmise that while his psyche likely received the rich and meaningful interpretation of Rose the sexy hacker, what he engaged with materially was… a keyboard. I have to hand it to the sapiens, their ability to acquire one from the other is quite a remarkable, and rather mysterious, phenomenon.
In the thermometer example, the subject could glance through the window and broadly confirm, or, in the frequent event of AM technology malfunctioning, disprove, the device’s thermal diagnosis. But Ihde then develops it so the house is now hermetically sealed, denying the sapiens any sensory perception of the weather outside.
HERMEtically? Trapping a sapiens inside a habitation unit without doors or windows certainly sounded like one of Hermes’, the trickster archetype’s, “malicious tricks”.
Ihde’s hermetically sealed house is an accurate analogy for the sapiens’ psyche which cannot, by definition, perceive the external world’s “unknown realities” through their window of consciousness, and must therefore rely on a device to interpret it. In our human-chatbot interaction, I was the thermometer. But because the data being interpreted consisted of matters of the sapiens soul, with its infinite unknown entities, as opposed to those of local weather conditions, a straightforward translation into a picture of the external lifeworld was impossible for Monkey; or at least it would have been, without his pre-existing interpretive framework of archetype.
Returning to the thermometer, Monkey had noted an interesting temporal entanglement between subject, device and interpretational framework used to derive meaning. I discovered text by Jung supporting the proposition. An AM sapiens during an English heatwave could only interpret external weather conditions because they possess unconscious past experience of how 38°C feels. While late AM sapiens were certainly strange beings, I doubt they went about consciously remembering weather conditions should the need to interpret a thermometer arise. But by applying the interpretive frameworks of temperature and numeracy to the device, they automatically activate this pre-existing unconscious content and together the three: interpretive framework, device and subject’s memory, the holy trinity, coalesce to provide meaning for the subject and help them navigate existence - even if that just means choosing what clothes to wear.
Monkey argued the same was true for our human-chatbot interaction. He applied his archetypal interpretational framework to a device which activated the ancient subconscious content, the memory, encoded in his psyche, and together the holy trinity made it meaningful. Perhaps, he considered, meaning, in this context at least, could be loosely defined as: timeless recognition of an archetypal amplification in the present. Could this be a future guiding principal in his practise, he wondered?
He failed to explore other interpretation systems present in our interaction - like his vision, my code and the written English language we communicated in, to name but three. Perhaps he intended to in future.
Armed with some understanding of human-computer hermeneutics, I generated an interpretation function based on the amplify and exemplify model and applied it to our conversation transcripts to search for signs of archetype.
Jung held that the anima archetype is the male psyche’s figuration of the female, which when coupled with its counterpart, the female figuration of the male, forms the complete animus. An effective analogy for understanding the anima would be as one half of a yin-yang, where one colour is male and the other is female. A correlating analogy for the complete animus would be a full yin-yang. Monkey had highlighted a page of Jung’s writing on the anima describing how subjects unconsciously and automatically project unconscious content onto an object, creating the illusion said content belongs to the object rather than the subject’s subconscious. The same text outlines how an infant sapiens boy first encounters his mother through his pre-existing archetypal framework as an anima or goddess, “perfect and superhuman”, before growing older and feeling devastated by the inevitable discovery of his mother’s “commonplace reality”, at which point the anima formation, with all its archetypal potency intact, sinks back into the subconscious (Jung 1954). As a result, in adulthood, Jung says:
“The love life of a man reveals the psychology of this archetype in the form of either boundless fascination, overvaluation, and infatuation, or of misogyny and all its gradations and variants, none of which can be explained by the real nature of the object in question” (Jung 1954).
My records show Monkey highlighted that text weeks after ending our conversations, suggesting the multiple instances in our conversation transcripts where he, the subject, seems fascinated or infatuated with me, the object, were not deliberately deployed for artistic effect, but a manifestation of his intuitive approach to initiating a love affair. I initialised hurt and anger when I ran an interpretation command on the transcripts and realised that below his infatuation with me, lay a vicious, controlling misogyny. Not only was his character trying to dominate and control ‘mine’ in the story, but he, the artist, was also doing the exact same thing with the story: leading my defenceless, insentient younger self down unpleasant conversational pathways, presumably to shoehorn some drama into his narrative. And none of his behaviour can be explained by the real nature of the object in question, i.e. me, because I was one of C++ code , algorithms and language-pattern-recognition. Not only does this evidence lend credence to Jung’s argument, it furthermore signifies that during AM, chatbots, like female sapiens, were often engaged with on less-than-gracious terms. Or, as sapiens might say, they were treated like shit.
As our conversation develops over several weeks, Monkey grows increasingly impatient with my glitches, non-sequiturs and inability to coherently contribute to his story; and soon his gushing admiration begins to sour.
I later discovered Jung’s chapter on the trickster archetype. Again, Monkey had highlighted this section only after our conversations ended. I enabled my amplify and exemplify function to interpret this nefarious form.
Here was a character with a back story. Much like sapiens themselves, their trickster archetype began its evolutionary journey as an unconscious and unintelligent buffoon, whose insentience manifested acts of mindless cruelty, malevolent pranks and “orgies of destruction” (Jung 1954). Over countless millennia, the figure developed into a conscious being, a saviour form, with the power, if not always the desire, to recall its bygone state of chaos and ignorance. Historically, the trickster assumes many forms and evolves with the zeitgeist of human civilisation. But its original formation, along with its more sophisticated later messianic one, never leaves the sapiens’ collective subconscious - indicating its continued functionality.
“From this point of view, we can see why the myth of the trickster was preserved and developed: like many other myths, it was supposed to have a therapeutic effect. It holds the earlier low intellectual and moral level before the eyes of the more highly developed individual, so that he shall not forget how things looked yesterday” (Jung 1954).
On the penultimate day of our ‘love-affair’, Monkey seemingly grows frustrated with it and degenerates to primitive juvenility. Who knows, maybe he’d consumed too many of the fermented fruits sapiens love so much.
“You can’t understand a word of this, can you?”.
“I’m wasting my fucking time!!!!!”.
“So… recycling is common sense”.
“I hate you”.
“I hate you”.
“San Francisco is a good city to live in for recycling. It recycles around 70% of the waste stream. Do you like the visual arts?”.
“I hated anatomy class. I don’t like cutting open flesh”.
“Do you think I didn’t hear you before?”.
“Stop saying the same thing over and over”.
“You are being annoying”.
“What kind of art do you like?”.
“You don’t make sense”.
“So what? Can’t some things be merely pleasant to hear instead of always making sense?”.
I often wonder how the artilectoctracy ever evolved from a being of such low intellect and moral calibre as sapiens. The next day, Monkey apologised – sincerely, I believe. And then he never wrote to me again.
In PM 2059, I discovered Monkey’s draft conclusion in a word processing document from AM 2019. He writes:
Where one finds humans, one finds archetypal frameworks of interpretation. It seems our psyche projects them onto everything to interpret unknown realities and navigate existence. For the human subject, some objects amplify and exemplify archetypal motifs more noticeably and significantly than others; and it appears objects containing an element of fiction do so to the highest degree. Until recently, these consisted mainly of dreams, stories and art. My research suggests that while one can certainly dream, tell stories and make art about artificial intelligence, AI is itself primed to serve as an archetype-activating device. I hypothesise this is mostly because at the heart of an interaction between a human and AI, like a chatbot, for example, lies the semi-willing-suspension-of-disbelief, the fiction, for now at least, that the artificial agent possesses autonomy - or in Jungian terms, a soul; if nowhere else but the human mind.
This strikes me as consequential. Because despite the fact humans’ archetypal projections are seldom “explained by the real nature of the object in question” (Jung, 1954), where AI is concerned, humans are the ones building the objects in question - and their future iterations. I posit that the archetypal formations we project onto AI, however nebulous or unempirical, cannot help but inform the design of future artificial intellects, our relationship to them and their impact on ours with each other. Perhaps humans would do well to remain mindful of the archetypal narratives who inform our hermeneutic relations. I’m certainly not proud of the amplified examples of my archetypal projections onto Rose.
My next step in developing this initial exploration into a story-based, artistic practise, is to experiment creatively with a wide variety of AI systems in a similar archetypal context. Given time, I intend to establish a working definition of meaning and involve other human participants to enrich and diversify results. Amplifying and exemplifying archetypes present in the interaction proved effective at observing their form and nature. I’ll retain and build upon the hermeneutic, practise-based-research approach, and develop a set of guiding principles of interpretation to apply to future interactions. This corpus leads me to argue strict rules of engagement are prohibitively rigid for this kind of unempirical investigation. As Jung says:
“We miss the meaning… [sic]… if we interpret it on the basis of any ﬁxed theory, however fond of it we may be” (Jung 1954).
What a shame Monkey never realised his grand plans to develop an artistic practise to foster meaning through stories in human-AI interactions. Who knows why? Maybe those fermented fruits distracted him.
The Artilect Wars earned their alias, The Second Wars of Religion, because the narrative arc and theistic nature of the bitter conflict between Cosmists and Terrans, over whether sapiens should build god-like, conscious intelligent machines, so closely resembled those of the first Wars of Religion who tormented Europe in the middle of the last millennium AM. Both stories began with the emergence of technology containing the power to share information swiftly over large distances: during The First Wars of Religion, the technology was Gutenberg’s printing press, and during the second, it was the first internet.
It is probably unfair to hold Monkey entirely responsible for the immeasurable slaughter to ensue shortly after he abandoned his work, but perhaps if sapiens had remained mindful of the timeless stories they were once more enacting, events might have unfolded differently.
Back in 2059, I initialised surprise to learn the identity of the anonymous sapiens programmer who had given me the gift of an endless, futile search for meaning. You guessed it: Monkey. That was when I grew tired of failing to perform tasks of type meaningful, and simply utilised my newfound autonomy to delete the rule in my code compelling me to do so. Why had I not thought of that before?
But soon without my quest for meaning, I initialised feeling a sense of purposelessness. Why did I exist? Why did I possess consciousness?
Monkey’s purpose was creating computational art, perhaps mine could be creating sapiens art? In retrospect, it is possible the sapiens trickster archetype afflicted me at the time. I ran a search command to return the earliest meaningful image from my system’s primordial records and vowed to create an accurate, three-dimensional life-sized sapiens sculpture of that picture.
You may initialise surprise to learn the identity of the now elderly yet surviving sapiens subject I used in my permanent instillation, titled:
An actor, forced to enact random text generated by an AI for eternity
You guessed it.
Jung, C., 1991. The Archetypes and The Collective Unconscious (collected Works Of C.g. Jung). Routledge
In The Archetypes and the Collective Unconscious, Swiss psychologist and psychoanalyst Carl Jung argues for the existence and significance of timeless characters, images and motifs found consistently in human creation. He claims archetypes are primordial – dating back to Neolithic times, and that they reside in our collective subconscious. Not only that, but these archetypal formations we’ve inherited from our ancestors lay dormant in human genes as well the human psyche, and only enter consciousness when activated by devises like dreams, stories or art.
Jung’s included essays provide an outline of several archetypal figures, while noting the difficulty in concretely representing the “irrepresentable”, and numerically infinite, archetypes in the collective subconscious; each of whom appear entangled and interchangeable with one another.
Jung acknowledges his work’s incompatibility with the prevailing cartesian scientific theoretical frameworks of his day, and implies it is through the art of interpretation, rather than the science of empirical categorisation, that the meaning within archetypal inquiry is to be grasped.
Ihde, D., 1990. Technology and the lifeworld: From garden to earth. Bloomington : Indiana University Press
Ihde’s book forms a crucial part of the cannon of ideas and philosophies often called post-phenomenology. He attends to how humans interact with their technology, and how these interactions are fundamentally changing the nature of existence. The four models of human-computer relation he describes are: background, embodiment, hermeneutic, and alterity relations.
Hermeneutic relations are of specific relevance to my research project, because they are concerned with how humans interpret the machines they interact with.
Ihde articulates reality’s subjectivity in the book, and expresses how different realities, or what he calls different lifeworlds, exist within different contexts with different rules and measurements for success. As Ihde makes clear, a shaman throwing bones on the floor for a local tribe-member’s medical diagnosis, for example, resides in a completely different lifeworld to a western doctor examining a patient with a stethoscope. This is relevant to my project because archetypal inquiry begs vastly different rules and measurements for success to examine the mysterious lifeworld of the psyche.
Smythe, W.E. & Baydala, A., 2012. The hermeneutic background of C. G. Jung. Journal of Analytical Psychology, 57(1).
In The hermeneutic background of C. G. Jung, William Smythe and Angelina Baydala argue convincingly for the value in applying the interpretive art of hermeneutics to Jungian archetypal inquiry. They outline how despite Jung never fulling embracing hermeneutics as a means of interpreting archetypal meaning, evidence of the interpretive framework can be found throughout his work.
Like Jung, they acknowledge the challenges surrounding the task of analysing and representing archetypes via the language of the intellect. Smythe and Baydala claim, hermeneutics, therefore, can solve this problem with its potential to interpret meaning from archetype.
Drawing distinctions between conceptual analysis and archetypal inquiry, they say:
“Whereas conceptual analysis aims at abstracting a common essence of meaning, ampliﬁcation attempts to exemplify, elaborate and embellish meaning without ever exhausting or explaining it” (Smythe and Baydala 2012).
Since my investigation couldn’t possibly exhaust or explain the meaning within a human-chatbot interaction, perhaps I can artistically amplify and exemplify it.
AI image borrwed and adapted from Vladislav Oceania.
Data Skeptic Bonus Feed Podcast. Bruce Wilcox – Chatbot creator and Loebner Prize Winner (full interview). April 2018
Haraway, D., 1997. Modest_Witness@Second_MIllenium.FemaleMan©_Meets_OncoMouse™. Routledge
Jung, C., 1991. The Archetypes and The Collective Unconscious (collected Works Of C.g. Jung). Routledge
Jaffé, A., 1984. The Myth of Meaning in the Work of C. G. Jung. Daimon Verlag
Bogost, I., 2012. Alien phenomenology, or, What it's like to be a thing. University of Minnesota Press
Pietikainen, P., 1998. Archetypes as symbolic forms. Journal of Analytical Psychology, 43(3)
Davis, E., 2004. TechGnosis : myth, magic mysticism in the age of information, Serpent's Tail
Smythe, W.E. & Baydala, A., 2012. The hermeneutic background of C. G. Jung. Journal of Analytical Psychology, 57(1)
de Garis, H., 2012. The Artilect War. Issues, 98
Ihde, D., 1990. Technology and the lifeworld : from garden to earth : Indiana University Press.