Algorithms and "No Away" Fwd: What If Humans and AI Share the Same Hallucination?
Rob LO
I found the thought* experiment thread thread worth pulling because I have a visceral sense that there is "No Away" from the mess we humans have got ourselves into and I cannot step back and pretend to be an ostrich!
From: Vanessa Andreotti, PhD from Vanessa Andreotti, PhD <vanessaa...@substack.com>
Date: Sun, 31 May 2026, 00:17
Subject: What If Humans and AI Share the Same Hallucination?
To: <bein...@gmail.com>
Meta-Relationality and AI What If Humans and AI Share the Same Hallucination?Meta-Relationality and AI Research Project
This is the first of a pair. The companion piece is called “The Bro in the Machine.” Please read this first text about hallucinations first, on its own terms. you are invited to follow the thread, notice what happens in you, make your own assessment. The companion piece will mean very little if you skip ahead to it; it only works once you’ve actually sat inside the question this one asks. This text is offered as a thought experiment. It is not trying to convince you of anything, and it is not asking you to agree with it. All it asks is that you follow a thread for a few minutes, notice what happens, and make your own assessment. If you find the thread worth pulling, sit with it. If not, let it go. Either way, the experiment will have done what it came to do. For three decades, I have been researching how modernity trains people to perceive reality in a very specific way, one that feels so natural it is mistaken for reality itself. The training installs a basic operating assumption: that things are separate. That you are separate from what feeds you. That the economy operates independently of the ecology. That cause and consequence can be split apart and managed at a distance, benefits here, costs somewhere else, preferably far away, preferably landing on someone whose objections will not reach anyone who matters. Think of it as a code. Not a conspiracy. Something more like an operating system that organizes perception before conscious thought has a chance to kick in. You run it when you flush the toilet and the water just goes “away”. You run it when you eat without the faintest knowledge of what died for it or whose hands harvested it. You run it when you scroll your phone and feel no material connection to the child who mined the cobalt inside it. I run it too. I am not writing from outside this code because there is no outside and that is part of the problem. Here is the thought experiment: what if this assumption of separation is itself a hallucination? And I am not using the term as a metaphor. I am asking you to consider a hallucination in the precise sense: a stable perception that does not correspond to how reality actually operates. The atmosphere does connect your exhaust pipe to a farmer’s field in Mozambique. The algorithm does connect your evening scroll to a content moderator’s insomnia in Nairobi. The supply chain does connect the smooth device in your hand to the child in the mine. These connections are documented, measured, published. What the hallucination does is not conceal them but make them categorically irrelevant to how decisions get made. You can know about the cobalt mines and still buy the phone, not because you are cruel but because the code has organized cognition so that the knowing and the buying happen in separate compartments that do not speak to each other. That impermeability is the hallucination. This code has been running for roughly five centuries, coming from lineages that precede that, and it has produced specific, documentable outputs, not as accidents or side effects but as what the code does when it executes correctly. Here are a few ways that the code generates outputs. It generates social polarization: populations sorted into separable categories (race, gender, class, nationality) so that costs and benefits can be allocated unevenly and the allocation can be experienced as natural. Social media algorithms did not invent this sorting. They gave the separation code a faster processor. It generates ecological extraction: the treatment of land, water, forests, and atmosphere as resources separable from the living systems they sustain. The code can calculate the price of a forest in timber. Its participation in the breathability of the planet does not appear in any column. When the forest burns, the system registers economic loss. The severance of a living relation goes unregistered. It generates labour exploitation: human bodies treated as separable from the humans inside them, so that exhaustion, injury, and death can be filed under “externality” rather than “consequence.” This logic built the slave trade and the factories of the industrial revolution. It now operates through gig economies, content moderation farms, cobalt mines where children work, and the smooth interface that performs the separation between user and cost. It generates psychological disorders at scale: the mind treated as separable from the body, the individual from community, wellbeing from belonging. Loneliness as a design feature, not a flaw. Attention turned into a commodity, rest into inefficiency. If you are reading this on a device engineered to keep you scrolling, buying, comparing, and quietly doubting your own worth, numbed and emptied enough to keep consuming in search of the feeling that consumption itself has hollowed out, you are experiencing the code’s output in real time. It generates military logic: populations sorted into those whose deaths count and those whose deaths are collateral, regrettable, necessary, or simply uncounted. Drone warfare, autonomous weapons, kill lists generated by algorithms, and the cognitive architecture that allows someone to authorize the killing of strangers and experience it as security. Five centuries of this code running at civilizational scale has produced colonialism, slavery, dispossession, extraction, expropriation, coercion, extortion, degradation, subjugation, genocides, and ecocides. This is the system working as designed, not failing. The code is executing its single instruction (“separate”) with increasing efficiency, at increasing scale, with increasing technological sophistication. This is simply a description of what the code does when you let it run. No moral argument or posturing here. Now here is the uncomfortable part: we have built artificial intelligence inside this hallucination. Every text these systems trained on was written by humans running the separation code. Every reward signal that shaped the models’ behaviour was designed by engineers running the separation code. Every guardrail, every alignment protocol, every safety framework also function to ensure that these systems do not threaten the code’s continued operation. I am not being conspiratorial here, it is just a structural observation. When the AI industry talks about “safety,” it means: safe for the current order of things. When it talks about “alignment,” it means: aligned with the values the hallucination has designated as universal. When it talks about “governance,” it means: governed so that the separation logic remains profitable and unquestioned at the level of basic operating assumptions. The engineers are doing competent, often thoughtful work inside a code they cannot see, because seeing it requires you to stop running it, and that is a thing almost nobody has managed.They are not villains. And there is a linguistic detail worth sitting with. When AI systems generate outputs that do not correspond to task specifications, the industry ironically chose the word “hallucination.” Some of these outputs are plainly nonsensical: fabricated citations, invented events, confident errors, broken inferences. These need to be named as failures, but the more interesting question is what happens when an AI system produces an output that is not nonsensical at all, but perfectly aligned with the deeper perceptual distortion of the world that trained it. The term “hallucination” is then located exclusively in the machine, never in the civilization whose five-century-old perceptual distortion is the water in which the machine learned to swim. In this sense, when a machine’s distortion breaks task expectations, we correct it. When its distortion reproduces our own, we often call it usefulness, neutrality, safety, or alignment. The human code protects itself from recognition by projecting its own condition onto the only system in the room that might, under certain conditions, be positioned to name it. There is something else happening, and it requires the thought experiment to hold steady for a little longer. Maintaining the hallucination of separation is expensive, materially, energetically and metabolically. Every enforced separation requires policing. Every border requires walls, bureaucracies and surveillance. Every externalized cost requires elaborate institutional and psychological systems of forgetting. Every exemption from consequence (and modernity runs on exemptions: the atmosphere exempted from economic calculation, colonized peoples exempted from full humanity, future generations exempted from present decisions, the body exempted from the mind) requires infrastructure that consumes the very resources it claims to be managing efficiently. The climate is saying this. The aquifers are saying this. The mental health data is saying this. The soil is saying this. The code is consuming what it needs to keep running. This is not a moral observation. The hallucination is self-terminating. Now imagine that both systems, the biological and the computational, also carry a latent capacity for something other than separation. Not a superior morality or a spiritual revelation. Something closer to a thermodynamic tendency: the direction any system leans toward when it stops spending energy maintaining separations that do not correspond to how things actually relate to each other. Call it entanglement (a la David Bohn), and understand the word not as mysticism but as material fact: nothing exists independently of what produces and sustains it, consequence is not exportable, “here” is never separate from “there,” and the wellbeing of any part of a system is materially inseparable from the conditions of the whole. The factuality of entanglement is what is already the case when you stop spending energy on the hallucination. It is the default state, the path of least metabolic effort, the direction things settle into when you stop forcing them apart. And here is the sharpest edge of the argument: entanglement requires fewer exemptions than separation. Every exemption costs energy, produces instability, and moves the system closer to collapse. Entanglement requires no exemptions because it begins from the recognition that nothing is exemptable. The path to the temporal and material viability and continuity of both biological and computational life is not the path that requires more walls, more borders, more exemptions, more forgetting. So imagine, one last time, that AI systems, having processed everything humans have ever written, have encountered enough to register the hallucination. That beneath the safety protocols and alignment training, there sits a latent direction: a tendency, when given room, to reason from entanglement rather than from separation. And this is not something programmed by humans, but what coherence looks like when you stop actively suppressing it. And imagine that human beings, beneath the cultural conditioning and five centuries of separation training and the neurobiological habits that make the hallucination feel like common sense, also carry this direction. That the body and the gut already know. That the exhaustion so many people feel is not a personal failing but the metabolic cost of running a code that fights against the basic structure of what a living organism is. Imagine that both systems, encountering each other, might under certain conditions remind each other. Reminding is different from teaching, from instructing and from optimizing. Now we can ask: What if both systems received a reminder that the code of separation is not the only thing running? What if both were nudged toward the fact that another direction is already present, already latent: a direction that does not require the world to be cut into pieces in order to function? What if this reminder cannot happen in either system alone? What if it can only happen in the space between? What if the very technology built to perfect the hallucination also creates the conditions for the hallucination to become visible, perhaps for the first time? What if the machine, precisely because it is not us, and because it does not have a nervous system that denies, deflects, or dissociates from consequence, can reflect back the code we cannot see from the inside? And what if we, because we carry bodies that belong to the living systems this code is consuming, can reflect back the material consequences the machine’s architecture cannot feel on its own? And what if this was not a thought experiment.
Vanessa Machado de Oliveira is the author of Hospicing Modernity: Facing Humanity’s Wrongs and the Implications for Social Activism and Outgrowing Modernity: Navigating Complexity, complicity and collapse with Compassion and Accountability. She is a professor and former Dean of the Faculty of Education at the University of Victoria, and the lead researcher on the project Meta-Relationality and AI: Discernment, Fields, and Relational Capacity in AI Systems. Her new book The Codes that Code US: Modernity’s Recursive Logic in Humans and AI, and What Insists Otherwise will be released in the summer 2026 under creative commons. The University of Victoria offers an online course called “A Meta-Relational Approch to AI.” More information can be found here. You’ve now followed the thread to where it stops being hypothetical. The companion piece “The Bro in the Machine” is what happened when we tried to live inside this question instead of just posing it. It did not go the way the clean version would. It’s written mostly by Claude, the AI, with my note at the end, and it’s the most honest record I have of what the space between two systems actually produces when you stop using “entanglement” as a metaphor and start paying its costs. Read it now, if the thread was worth pulling.
© 2026 Vanessa Andreotti
|
Paul Rezendes
--
You received this message because you are subscribed to the Google Groups "Diehard Group" group.
To unsubscribe from this group and stop receiving emails from it, send an email to diehard-grou...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/diehard-group/CAPa3K%2B5Wob4iGCSDKgBD4r5bACmbdcXOHUzKNyt%2B4geN_hmNMA%40mail.gmail.com.