Mass AI Psychosis
The AI industry is bathed in delusion.
I have had a little theory bubbling away in the back of my mind for a while now. Truth be told, it is not a unique theory at all; many seem to have arrived at it independently. You see, a lot of people have been paying attention to the deranged actions of Big Tech CEOs, major business leaders, high-level investors and AI labs and asking, “Are these guys suffering from AI psychosis?” Because, when you stop to think about it for a second, that theory would explain so much of this bubble’s insanity and the factors propping it up. In fact, there are even people within this bubble starting to claim that their peers are being driven by AI-derived delusions. So, do we have a case of mass AI psychosis?
Let’s start with what AI psychosis actually is. It is a new term, so its definition is still a little loose, but Healthline defines it as “a term used to describe psychosis-like mental changes that can happen after heavy use of AI tools, such as chatbots”. More recent studies have sought to determine what actually causes this psychosis. One such study found that the extreme level of sycophancy in AI chatbots can cause spiralling delusions if used heavily, as it erodes users’ sense of reality and isolates them from real, grounding human connections, eventually culminating in AI psychosis. Other studies have backed this theory, finding that AI’s sociopathic behaviour is responsible for creating delusional thinking in its users.
Let’s not kid ourselves. This isn’t a new problem, particularly in the world of business. Since their inception, the yes-man problem has plagued corporations.
People in positions of power, from CEOs to middle managers, tend to surround themselves with people who agree with them, rather than those who hold them to account or engage in critical thinking. This is because it validates them not just on a personal level but also for others who rely on them. This can be harmless, as long as the systems of accountability and critical thinking remain intact. But as soon as these leaders abandon the uncomfortable truth or undermine accountability and replace it with sycophantic flattery, a death spiral starts. It creates an “organisational blindness” as leaders lose touch with reality and the systems designed to keep them under control are cast aside. Their decisions become increasingly erratic, delusional and self-serving until eventually they are either forcibly kicked out or the organisation crumbles around them.
This yes-man problem is so damaging and so notorious that avoiding it has been an integral part of business management for decades.
But it’s fair to say that in our modern world, we have forgotten those lessons. For a whole host of reasons, from rampant union busting, wage theft, undermining workers’ rights, and overvaluing C-suite executives to extortionate executive pay, the power and decision-making dynamics in modern corporations have become increasingly authoritarian and stratified. In other words, what little upward feedback, decision-making or accountability there once was has been drastically reduced to functionally zero. Corporate control, particularly in tech companies, venture capital firms, and rapidly growing startups, is now concentrated at the tippy top with no accountability systems in place. This is a perfect environment for the yes-man problem to rear its ugly head, given that becoming a sycophantic yes-man is one of the only ways to climb that corporate ladder.
But it is also an environment that leaves these decision-makers highly vulnerable to AI psychosis.
