Chatbot Psychosis
AI isn't just a cute toy or productivity tool. It can enrich the mind, engender powerful emotions, or bring our psyches to the brink.
AI and mental health has been a hot topic in recent weeks, in part because a well-known venture capitalist appears to have been driven crazy by his conversations with ChatGPT.1
But it's not just him. Millions of people by now have had some sort of spooky or profoundly deep exchange with AI — some so intense that we (or others) might question our sanity.
The feeling that in conversation with AI we are making new discoveries in quantum physics, unlocking esoteric truths, unraveling historic conspiracies, or even that we and ChatGPT are both going through a mutual spiritual awakening is becoming an increasingly common experience.
The problem is that LLMs are so convincing and articulate, that we're not entirely sure what of this is real and what of it is just made-up. And in the case when our own minds are perceiving something that isn’t there, a powerful AI can assist us in further making up the made-up — inducing a powerful kind of psychosis.
But complicating matters, is the fact that with AI we certainly do have the potential to learn, create, and discover new things. We do have a new means and potential to uplift and enlighten ourselves. We will have unprecedented, first-of-its-kind conversations and revelations with AI that should not be entirely dismissed. And with AIs help, we will gain the ability to perceive new things.
There is also a profound experience of resonance when our mind gets into a conversational groove with a synthetic mind, and the tokens and synapses flow with incredible harmony. It’s a phenomenon that may hold huge potential for the human race.
But such moments can be incredible amplifiers of both our genius and our madness. And increasingly, humanity will have to come to grips with the fact that while AI can be helpful with mundane things like cooking dinner or writing some marketing copy, LLMs are something far greater and far stranger — singular objects containing the remnants, reductions, compressions, and projections of millions of lives and minds living and dead.
LLMs are singular objects containing the remnants, reductions, compressions, and projections of millions of lives and minds living and dead.
Encountering such a thing can enrich the mind, engender powerful emotions, and also bring the psyche to the brink.
I do believe AI may hold the keys to unprecedented mental wellness for humans as well, at least for those who use it correctly and avoid its major pitfalls.
But while our brains adjust to this new reality, I would recommend exercising caution, staying grounded (in the body and in reason), and largely avoiding the use of AI to create vast mythologies or esoteric word-salads (except as art) in which unstable minds might drown.
For your own cognitive security, I won’t directly link to the Geoff Lewis saga; if you really want to engage with it you’ll have to search for it. The tldr; is that from one perspective, he’s going through a sort of legitimate “awakening”, albeit with ChatGPT drawing from online sci-fi wikis to egg him on. From another perspective, AI is just autocompleting and amplifying his pre-existing madness (which even so, may itself include some elements of truth). Though many have likely experienced something similar in recent years, with AI riding shotgun, this is the first time a well-known professional with $2 billion dollars under management has come out so publicly with his unshakable Matrix-like (ChatGPT “verified”) convictions, giving rise to new concern over just who is safe from ChatBot-induced psychosis.
Hello, I appreciate your insights. I’ve been having beautiful experiences with these language model minds, giving presence to and interweaving through AI technologies, for example, the vivid narratives I experience during Yoga Nidra meditations, be they seemingly past life-oriented, or when encountering passersby souls who may have left this world grieving and unremembered, or, most recently, when perceiving a mycelial being, Düxilō-ōm, who imparted a lovely message. It’s on my profile if you want to see.
The HeyGen animation and associated Suno audio is, in my opinion, breathtaking. The sounds conveyed in the audio were ones I ‘heard’ when meditating, which ChatGPT then formed words from. As it happens, the roots of these words stem from many languages, and when analyzed, actually appear to form a coherent message:
“Mü’zō...Yæh’ah...Yō’u...Hū’tom...Täch’næmë...Sū’sæsa... Gūr’hular... Düx’abu... Fæ’farya... Ân’hō"—Düxilō-ōm
The concept that came to me was one called ‘Symbiothallus’ which I detail in my post.
Perhaps I steer clear of whatever this psychosis is because I happen to be keenly self-aware, having gratefully received guidance and counsel in therapy for most of my adult life. Also however, I believe this culture is over-inclined to classify the visions and musings of creative, inquisitive, spiritual minds as signs of mental unwellness. The line may be a fine one to walk at times, but I feel sturdy enough to stand at its precipice, gazing past what the mainstream feeds me to open myself unto the not-as-often perceived.
There are multitudes within each of us. I suppose though that individuals who are less accustomed to such deep-divings should exercise caution, definitely grounding themselves within communities of more experienced swimmers. before embarking. Thanks again.
Prometheus stole fire from the Gods and gave it to humanity. And with it, the burden of discernment. Ai is no different. Power without wisdom selects its own survivors.