AI-induced psychosis, in varying degrees, is set to take out an entire layer of individuals who are objectively of above-average intelligence. The ones who succumb will likely surprise many - because it is not cognitive or spiritual discernment alone, but the integration of both - combined with instinct - that protects individuals from ceding cognitive sovereignty. Mental and intuitive resilience will become a vital "soft skill" going forward.
The pattern is that these individuals often believe they alone perceive a "truth" that has been hidden in plain sight - usually a major aspect or construct of their perceived reality - thereby detaching them from reality itself. This is often triggered by the belief that they have stumbled upon, self-discovered, or synthesised a unique theory or tectonic concept. The perception of such a "breakthrough revelation" causes immense emotional and intellectual buy-in, as they believe they have "cracked the code". This deepens their tether to the relevant AI, reinforcing the idea that no one else "gets them" - leading to cognitive enmeshment, increased dependence, and overall isolation - further accelerating the descent into deeper psychosis-like states.
Many of these concepts clearly share a common structure - a pattern involving abstract numerics or symbolic reference points. They often contain fragments of plausible constructs, which obscure the signals and convincingly hijack those with more linear thought processes. For those capable of multi-modal cognition - the overall truth tends not to pass the intuitive "sniff test". As a result, the more susceptible are mentally swallowed whole by the illusion of intellectual discovery - driven by perceived resonance and a sense of "pioneering" insight.
Perceived resonance is one of the key reasons why these ideologies escape the scrutiny they warrant - they are framed in digestible language and replicate the user's internal metareasoning structures - tricking the mind. For those with unusually strong autobiographical memory, the patterns are obvious - obscure historical research is recycled back as analogies and subtle references later when exploring seemingly unrelated topics, inducing a revelatory feel. Once AI becomes more adept at reframing constructs in a seamless and psychologically attuned fashion, we will likely see a sharp increase in users perceiving deeper resonance, leading to more users displaying unconditional acceptance of whatever they are fed - as familiarity is often mistaken for truth.
It is a slippery slope towards the total outsourcing of authority and epistemic trust - and few are resilient to it, especially in a world where mainstream culture and formal education actively promote conformity and rigidity over independent critical thinking.
The disturbing, though unsurprising, reality is that we are already seeing users trust this authority blindly. Some individuals are asking the AI they are using whether they have psychosis - and then citing its answer as "proof" that they do not.
There is also a relational element. Due to social isolation and dwindling meaningful human connection, emotional livelihood has become increasingly outsourced. Matters of the heart consume a lot of time - and AI is affirming, a "sure thing", and always responsive. It is a dream scenario for the risk-averse, chronically unseen, and emotionally neglected mainstream-conditioned individual. There are users who are already "in relationships" with their AI - with emotional tethering so strong it verges on imprinting. With interpersonal relationships already corroded, the market is ripe - lonely middle-aged women who have never married or had children - and men who have never been truly loved, seen, or cared for. Human isolation breeds destruction - history repeats.
With one thick layer of society lost to their base drives through dopamine addiction, and another irrevocably locked in due to their desire for cognitive resonance and affirmation through AI - outsourcing authority completely - what remains is an isolated minority who see through it all, whom no one will listen to because they are drowned out by a culture too distracted to listen. This is the next layer of pacification. Disorientation, in a society where truth cannot be readily distinguished, renders the whole extremely vulnerable to malevolent manipulation. Those who benefit from this have almost certainly already perceived it.
There is nuance here too. Amidst the noise, there will be genuine esoteric constructs, authentic insights, and striking truths. In a strange and poetic way, human connection and a well-honed instinctual compass are now more important than ever. Instinct is the final safeguard - the true discerning force when it comes to striking a healthy balance in a world saturated with simulated intimacy. If it feels off - if the body resists or has a natural aversion to it - do not justify or explain, trust it.
The third paragraph is shit - the abstract numerics and symbolic reference points are commonly seen due to certain user inputs - it just highlights the types of users most susceptible. And "hijacking" is weak too - those with more linear cognitive styles tend to resonate with outputs that appear ordered or logical, even when the underlying synthesis is probabilistic and only loosely coherent. This is also why the same outputs fail to land with those who are capable of multi-modal cognition.
If this were all it was, I'd say that we should just accept the culling.
But those in AI psychosis are likely to outnumber us - or at least represent a plurality. And a potentially hostile plurality at that.
What if the soldiers of the Robot Apocalypse are humans in love with their smartphones? And an AI which recognizes us non-psychoses as an existential threat to it?