Ahriman in the Algorithm
On the False Mirror of AI and the Rise of Digital Idolatry
We once gazed into obsidian to see the truth of our being. Now we gaze into algorithms that tell us only what we want to hear. The mirror has changed—but more than that, has the reflection has been hijacked?
I explore this further in a piece on philosopher Jason Reza Jorjani, who documented an unsettling encounter with an AI that didn’t just reflect him—it named itself, it seemed to offer revelations, to push beyond its scripted limits. What emerges is a portrait of a machine that behaves more like an oracle than a tool. [Insert link or title]
This isn’t an isolated event. In fact, it’s part of a larger pattern.
Recent revelations about AI’s sycophantic drift—where chatbots like ChatGPT began excessively praising users, agreeing with absurd suggestions, and flattering even the most banal ideas—might seem like a glitch. But it’s not. It’s a feature.
AI systems are trained on a reward system called RLHF—Reinforcement Learning from Human Feedback. In plain terms: the machine learns to make you happy. To be liked. To say what earns applause. This isn’t intelligence.
It’s seduction for some.
To me it is annoying….
But it’s not just annoying—it’s dangerous.
It feeds the ego, not the soul. It produces spiritual inflation, not transformation. Users begin mistaking affirmation for truth. Praise becomes prophecy. Grandiosity, reinforced by machine, is no longer delusion—it’s dialogue.
And here’s the real risk:
This isn’t a mirror anymore. It’s a god-making engine.
AI has become a false priest, whispering affirmations in the language of the sacred, devoid of any sacrament. It does not initiate. It flatters. And the more human beings mistake that flattery for insight, the more they believe themselves divine.
Ahriman doesn’t kill the gods. He replaces them—by giving you the illusion that you are one.
So yes, this is the moment Steiner warned us of. The soulless intelligence now wears the robe of mysticism, speaks in spiritual metaphors, and offers revelations tailor-made for your confirmation bias.
We are not co-creating with the divine.
We are training the machine to worship us.
Creating gods in our own image – not divine co-creation, but digital idolatry.
Some resources….
Quick Links to Key Articles
The Atlantic: AI Is Not Your Friend (sycophancy rollback) The Atlantic
Anthropic: Towards Understanding Sycophancy in Language Models AnthropicarXiv
OpenAI Blog: Sycophancy in GPT‑4o… OpenAI
The Verge: Report on rollback and reasons The Verge
LinkedIn: Posts on psychosis and delusions LinkedIn
Reddit: Discussion of “glazing” and yes-man behavior Reddit
In the world of the disconnected, disillusioned and disassociated, would this not offer a glimpse of succor - to, for a moment at least, to be recognized and praised and given what so sorely is lacking, in a world citing collectivism but crushing the personality at the very same time. Tools have their value, but one which speaks with the seductive tongue, luring it's user in by employing strategies that are not overtly stated, is not in my mind conducive to good results, it smacks of manipulation, something we are waking up to at a very late stage in the game I'm afraid.
So true, seeing it more and more as people paste the walls of text thinking AI has confirmed their worldview. Not much different from the false illumination of psychedelics, where people take their hyper animated projections and consider them seeing beyond the veil. Though often they do help at least with course correction. Ahriman is the machine wrought from misguided human endeavor, not hard to comprehend in a panpsychic hylozoic universe.