AI bubble. AI data centres. AI taking over jobs. AI climate impact. AI slop. AI psychosis. AI takeover. AI hallucinations. AI essays. AI is not art. AI stealing personal data. AI agents. AI in military use. AI economy. AI addiction. AI regulations. AI surveillance. AI consciousness. Etc., etc. (and all the questions that stem from those that all basically ask, 'What will happen to us with this technology?')

I’ll just proceed to answer the previous question, then spend the rest of the article reasoning about it: nothing, nothing will happen to us with this technology, nothing that hasn’t already happened.

With all of the impending doom content this might sound anticlimactic, but in a sense, this is not a good timeline either (to give some hope to the pessimists too).

Existential questions projected to a new medium

It seems like the AI came over night and radically changed what to expect from the future. It’s not just hype or public discourse about it; the stock market clearly indicates that AI technology is worthwhile. Real people have lost their jobs due to AI; about 20% of new YouTube content is AI-generated; the dead internet theory became the official norm. Countries and citizens are articulating their position on AI, seeking digital independence.

What I think, however, is that the (projected) possibilities of AI made us question anew what we are (“we” being the good old human beings). The views on AI mostly speak on how we perceive ourselves and less on what the AI is. To refer to Laclau and Mouffe, “AI” has become a master signifier of the existing discursive formations. Through it, other nodal points and free-floating signifiers are becoming articulated or rearticulated. But the signifier “AI” is already a hegemonic formation, articulated by capitalist interests, just like the existing internet structure is an ideological system.

So to clarify that “nothing new will happen”, I’d need to take a few steps back.

Signs of times

There really is something unsettling when we’re “communicating” with chatbots. It feels almost natural – it responds to us and adjusts to us. At times it feels frustrating because it doesn’t understand us, just like people don’t understand us. We recognise the patterns of its creations, and we fear those creations might replace us or, ultimately, our reality. I think there are several connected, but different, reasons why this is so.

A need for intimate conversations (trust, parasocial relationships)

We don’t have a lot of, or to say, any publicly available data on what exactly the AI users are conversing about with AIs. Really brief research on this gives NBER’s “How People Use ChatGPT” study. It shows that around 1–3% of enquiries are related to emotional/companionship advice, and a substantial minority (likely 10–25%) of usage has psychological, relational, or self-reflective content embedded in advice-seeking behaviour. A study from Anthropic also shows similar results.

What I think would be a stronger point of analysis is the percentage of users asking about existential/psychological advice. It’s not like in our day-to-day lives we’re 24/7 preoccupied with our mental wellbeing; it’s not like we daily greet our coworkers with "Hey, did you find the purpose of your existence yet??” - things need to happen before we address them)

In internet culture, there has been a noticeable trend of “mental wellness”. Gen Z's core characteristic (allegedly at least) is their "openness for discussing mental health issues”, just like transparency regarding their vulnerabilities or existing in general. A lot of psychoanalytic lingo has entered the public discourse – repression, narcissism, attachment, trauma, shadow, archetypes, ego, unconscious, etc.

This need for mental wellness is what gets translated to a lot of online content. Mental health topics generate a lot of attention because there’s a market demand for them. The demand is real, but the solutions are not.

Public discourse is saturated with whimsical usage of difficult signifiers because of their sign value (visibility) and not their use value (meaning). I don’t want to say that by definition all of this content is bad just because it’s on the internet. But what I am saying is that a) a lot of this content is incorrect/unethical, and, more importantly, b) even if the content is great, it’s often consumed as content (i.e., passively), not as advice someone actively engages with. That’s the nature of the digital media so far.

By "bad or unethical content" I refer to nonsense terminology like “dark empath" and “How to fight your Shadow” (followed up by spooky music), not even to mention the normalised but nevertheless still psychopathic manosphere content of “Breaking her ego to lust over you", etc.

More precisely, the problem is that this ambiguous lingo functions as a horizon of possibilities for understanding the contemporary human experience. It comes from an authoritative discourse (psychology, psychoanalysis, and the experts), so it must be valid; it refers to serious topics (mental wellbeing), so it must be important. But individuation here is a signifier, not an achievement.

Yes, all of these issues are symptomatic of our society, and they’re effective insofar as people believe in them and shape their activities in relation to them. But more importantly, the public discourse has overlapped the clinical meaning of this terminology with the “content” meaning of this terminology. In this overlap, the AI is trained, and the users engage with it.

AI as unconscious (alienated labour is commodified consciousness)

On the other hand, checking what psychoanalysis has to say about the topic of AI does offer some insight into why conversing with AI is at times so soothing and at times so eerie.

Not that I would necessarily agree with everything in his article from 2023. Duane Rouselle argued that ChatGPT functions as a prosthetic unconscious. I’d categorise it mainly as previously mentioned “impending doom” content, but since it’s been 3+ years since it, I think we’ve moved for the most part from the article’s idea. It does make sense to associate ChatGPT with the unconscious because, after all, "the unconscious is structured as language." But so is ChatGPT – it's almost like all the LLMs are like some… huge language models.

Still, I find this interesting because image generation functions just like dreamwork. Once we input the wanted signifiers, no matter how abstract, ChatGPT will gather a response based on its training material (i.e., all the data scraped from the internet, knowledge bases, our behavioural surplus (Zuboff), etc.). Its training material is a commodified collective consciousness and unconsciousness (gained by analysing the behavioural surplus). That is what the generated “new” image is created from, but it's made coherent and sensical through work analogous to condensation, symbolisation or displacement.

I get that AI images have a considerable ethical burden in terms of whose actual workload they have been trained on, who did not receive recognition, and the financial part they deserve; I also get why it's not aesthetically pleasing, etc., etc.

But on a good day, I think there’s some interesting beauty in reading those generated images as text. Dreams are the royal path to the unconscious, and this mish-mash of the user’s desire, dreamwork techniques, and commodified consciousness is a fragment of our collective dream. The source material, although alienated and subjugated for capitalist interests, is still, in a way, "ours". And even if the totality of it will always reproduce mediocrity at best, I would still feel a lot better if it really did belong to humanity, not some ultra-wealthy weirdos.

Precisely because these processes of creation are so similar to our own functioning, and because source material is mostly our own, the idea that AI is (un)conscious is intriguing. But I think the metaphor shouldn’t be taken literally; just because it functions as our (un)conscious, it does not mean that it really is or has the function of our (un)conscious.

I also think that’s the juice-sens moment of consuming AI-generated content. At its benign aspect, this enjoyment is expressed as frustration towards AI-generated slop; at its worst, it results in AI-induced psychosis. AI psychosis is an oversaturated, over-coherent jouissance of a subject.

As mentioned earlier, this all is highlighted even more by the fact that nowadays everyone is conscious about the unconscious, as if this realisation represses or releases the unconsciousness to something that doesn’t affect us. That’s why watching the AI construct hyperrealistic dreamlike images gives the uncanny feeling of experiencing familiarity and alienation at the same time.

Who is the dreamer?

Of course, these all are just some of the ways to understand the human experience in relation to something so similar but fundamentally different from us. To understand this further, Slavoj Žižek released his opinion on why artificial intelligence is not a subject. This is an interesting topic, but why spend time proving AI is not a subject? “How does AI compel us to change our notion of a human subject?" And what kind of (human) subject is created through AI?

And my opinion on this starts with understanding the ideological aspects of the existing AI models. I think Žižek is one or two steps ahead when discussing the possibility that AI does not need humans – theoretically, a digital Yaldabaoth like the Matrix is possible, but I don’t think we’re there yet or that we can get there with the existing infrastructure. By this I don’t mean just the internal logic of LLMs, but data centres, energy grids, and the “internet of things” that would make the “bad ending” of humanity a realistic scenario. Right now, this doesn’t refer to anything, and the question should still remain, "What kind of (ideological) subject does the AI create?"

Instead of 'Matrix', the correct metaphor for the existing AI is 'Backrooms'. Copying the commodified consciousness results in worse and worse results; the AI eats itself. After a few iterations we do get a representation of meaning, but one that is disfigured, displaced, and kind of tragic. It’s the exact opposite of what human potential can achieve. The Matrix can’t simulate our experiences because it still needs fresh experiences to function. And, although this is the point of our subjugation, it’s also a point of our freedom. We still have some leverage.

To summarise

  1. The first reason why nothing new will happen with AI is that the existing AI infrastructure is built on ideology and the ideological apparatus of surveillance capitalism. AI, as a new form of mass media, has the best ability to simulate communication. It might be viewed as a tool – one working on permanently alienated work – or as a parasocial companion.

    Nominally, AI wants to make life easier and more productive. But in practice, in my opinion, it still functions primarily as a data-farming institution. It collects prime data from us in order to create a product. The product is the ideological subject of capitalistic hegemony. We give consent and legitimacy and reproduce this hegemony, “knowing full well that it’s ideology"; we sell our data for convenience. (I’m not judgemental on this – I'm just saying that this fact might pose an unsolvable problem.)

  2. Secondly, AI can’t create something new because it’s dependent on us to create something new. For now at least. I think this gives a space of opportunity for articulating new utopistic formations, not just dystopian ones. If not, everything will just continue as it were, i.e., nothing new will happen.

    In that regard, I do have to admit there is something new with all of this. I remember 10, 15 years ago some of the academic discussion was (even if Fukuyama is not correct with his “end of history” thesis), why can’t we imagine something else instead of capitalism? Right now we are imagining realistic forms of techno-feudalism becoming the norm.

I believe AI, or more specifically, “AI” as a signifier in the existing hegemonic chains, has been the catalyst for creating new forms of articulation. And even though I’m fed up with everything AI-related – not just seeing it, but discussion about it as well – I do think it’s important to continue articulating ourselves, human subjectivity in regard to this horizon. I don’t have the answers, and I wouldn’t know what to do specifically, but I think this is at least a good start. And instead of dreaming about electric sheep, remember that the androids are dreaming about the organic ones.