Unlocking the Collective Unconscious

On Slavoj Žižek's "ChatGPT Tells What Our Unconscious Mind Radically Represses" (Google translated from Turkish)

Chatbots are machines of perversion, Slavoj Žižek writes, and cover up the unconscious more than anything else. An analysis of digital intelligence. ( April 2023 )

I told you about an accident that happened to me a while ago: A black friend of mine was so moved by something I said that he hugged me and shouted: “You can now call me ‘n…i’!”

One critic recently claimed that those who agree with me on this issue are “crazy”: “The problem is that Žižek’s arguments are based on the freedom to use racist expressions. Žižek uses the N-word as an argument against political correctness, thereby implying that black people not wanting to be racially insulted is politically correct, and therefore irrational. Perhaps the man he was talking to was not offended at all. But whether a non-black person says the word should not depend on whether they find a black person who ‘allows’ it. The use of words should be based on how the world is perceived. The N-word is a word used to directly legitimize one race’s ownership of another race. That’s what bothers me, man.”

It was an expression of friendship

Let me be clear: Like a chatbot, my critic ignores the explicit context of my example. I did not (and never would) use the N-word in any communication, and the black person who said to me, “You can now call me ‘n…i’!” did not mean that I should do so. It was an expression of friendship based on the ironic and friendly way black people sometimes use the word among themselves.

I am sure that if I actually called him “n…i” he would be annoyed at best, because he would think I was not getting the obvious point. This statement of his fits the logic of “an offer that must be refused”, which I have explained in detail elsewhere. For example: If I say, “What you just did for me is so nice that I don’t care if you kill me!” I certainly don’t expect the other person to say, “Okay!” and pull out a knife.

Chatbots' stupidity is their value

My guess is that chatbots, at least at the moment, can’t respond to such offers that should be rejected. (This is leaving aside the rare case where, in a very specific context, not just the use of the n-word by a person of color, but even the implication of it, can be hurtful. Similarly, the phrase “so help me God!” applies. If God were to actually intervene in my world at this point, I would be completely shocked.)

But have I relied too much on academic reactions to chatbots? Especially by deriding ChatGPT’s flaws and mistakes? Against this common view, Mark Murphy argues in a dialogue with Duane Rousselle that “AI is not a substitute for intelligence/sensitivity.”

According to Murphy, “The stupidity, the slip-ups, the mistakes, the irrational short-circuiting that chatbots do — the constant apologizing when they’re wrong — is their very value.” This allows us (as “real” people interacting with chatbots) to maintain a false distance from them, and when the chatbot spouts nonsense, we can say, “This isn’t me, this is an AI machine.”

ChatGPT is an unconscious

Rousselle and Murphy support their claim with a complex argumentation, and their starting point is the statement that “ChatGPT is an unconscious.” The new digital media externalize our unconscious in artificial intelligence machines, so that those who interact with artificial intelligence are no longer shared subjects subject to symbolic castration and therefore without access to the unconscious. As Jacques-Alain Miller puts it, we have entered a universalized state of psychosis with these new media because symbolic castration has become deactivated.

A horizontally divided subject has been replaced by a vertical parallelism – in fact, undivided – in which subjects and the externalized machine/digital unconscious stand side by side: narcissistic subjects exchange messages through their digital avatars. In this flat digital environment, there is no room for the “incomprehensible monstrosity of the neighbor.”

The Digital Unconscious: “An Unresponsible Unconscious”

The Freudian unconscious implies responsibility through the paradox of intense guilt that we feel without knowing what we are guilty of. The digital unconscious, on the contrary, is an “unresponsible unconscious” and this poses a threat to the social bond. A subject is not existentially involved in communication because this communication is carried out by an artificial intelligence, not by the subject itself.

“Just as we create an online avatar to join online groups and interact with others, can we similarly use AI personalities to take over risky functions when we’re tired? Like bots used to cheat in competitive online video games, or like a self-driving car can navigate a critical journey to our destination. We just sit back and encourage our digital AI, and it works until it says something completely unacceptable. That’s when we step in and say, ‘That wasn’t me! That was my AI.’”

According to Freud, Dreaming is the Royal Road to the Unconscious

Therefore, AI does not offer “a solution to the discrimination, the fundamental loneliness and antagonism that we still suffer from; for without responsibility, there can be no ‘post-givenness.’” Rousselle used the term “post-givenness” to describe “the realm of ambiguities and linguistic instability.” This realm makes it possible to approach the other in a state of supposed non-relation. It is directly a matter of the impossibility of how we relate to the other. It is about coping with the dark and incomprehensible monstrosity of the neighbor, which can never be destroyed even if we offer it the best conditions.

This "dark monstrosity of the neighbor" also concerns us, because our unconscious is a complex of dirty pleasures and obscenities, a dark other in the subject's innermost self. For Freud, the dream is the royal road to the unconscious. Therefore, failure to take into account the dark monstrosity of the subject also means, logically, failure to dream.

The Clownlike Feature of Père-verse-ity

“Today we dream outside ourselves, and that is why systems like ChatGPT and the Metaverse operate by offering the void we have lost, because the old castration models no longer work.” We are witnessing a direct intervention of the unconscious with the digitized unconscious – but why are we not then overwhelmed by an unbearable proximity of jouissance, as in the case of psychotics?

At this point, I would like to object to Murphy and Rousselle’s focus on how “pleasure can be delayed and denied” with AI machines: just as humans create something extremely obscene and avoid taking responsibility, a system is created that imitates the divided subject. This allows us to say, “This does not belong to me,” explicitly. Pleasure emerges at this point, in the refusal to be agential: we point to the AI ​​and say, “Look how stupid!” The clownishness of père-verse-ity lies largely in online conservatism’s need to resurrect the father. From Trump to various life coaches, many people act as prosthetic father figures. In these cases, we see the reactionary revival of a prosthetic phallic logic to sustain the idea of ​​“everything.” “(…) When a figure representing castration fails to emerge, we now see the direct invention of an unconscious without the structuring point of the father.”

The Perverse Return of the Obscene Father

What characterizes the AI ​​in this case is not a psychotic closure but, as Lacan puts it, perversion (or père-version, “version of the father”). The unconscious is not primarily the reality of jouissance repressed by the castrating father figure, but rather the castration of the father figure itself; this implies the absence of the symbolic function of the big Other.

The perverse return of the obscene father (Trump in politics, for example) is not the same as that of the psychotic paranoid figure. But why? With chatbots and other AI phenomena, we are faced with a reverse exclusion: to repeat Lacan’s classic formula, the excluded symbolic function (the Name of the Father) does not return in reality (as a representative of paranoid hallucination); instead, the dark monstrosity of the neighbor – that is, the impossibility of reaching out to another – returns in the symbolic, and this manifests itself in the “free” and smoothly functioning space of digital communication.

The Unconscious Is Repressed

This kind of reverse exclusion characterizes perversion, not psychosis – which means that when a chatbot produces obscene stupidities, it cannot simply be said that this is a case of me enjoying them without taking responsibility, because “it’s not me doing it, it’s the AI.” What happens here is rather a perverse form of denial: even though I know full well that the machine is doing the work, I can still enjoy it as my own production.

The most important feature to consider here is that perversion, far from making the unconscious (hitherto repressed) explicitly apparent: as Freud noted, the unconscious is nowhere as repressed and inaccessible as in perversion. Chatbots are machines of perversion, and they cover up the unconscious more than anything else: because they allow us to vent all our dirty fantasies and obscenities, they are therefore more repressive than even the strictest forms of symbolic censorship.