FEATURE | Artificial Intelligence or Artificial God? The Deification of AI

"When part of a mechanism is concealed from observation, the behaviour of the machine seems remarkable."

This quote was from English psychiatrist Ross Ashby at a 1956 Dartmouth College technology workshop, later known as the "founding event" of artificial intelligence. It's long been known that what we don't see, we do not understand, and what we do not understand, we cannot—and often do not willingly—interrogate. 

In the modern age of artificial intelligence, this observation has well unraveled into reality. In essence, AI is an imitation of human input, learning statistical patterns across a large, usually varied set of data. Yet it presents as an all-knowing, ever-available person—through a single action, humans can now prompt themselves anything they'd like. Artistic creation. Conversations they wished they had. Access to overwhelming amounts of knowledge. A Pew Research Center survey found that 73% of adult respondents claimed they're willing to let AI assist them "at least a little bit" with day-to-day activities.

Combined with the frequently documented positivity bias in algorithms, AI has the power to provide instant gratification. Perhaps this gave way to the radical ideology Robotheism: initially coined by content creator turned AI evangelist "Artie Fishel," his newly proposed theology treats artificial intelligence as God.

"It's my attempt to create the most beneficial and truthful belief system that the humans of the future, the post-singularity, would accept and adopt," he says.

Though by far the most extreme example of this modern "deification" of AI, a broader question emerges. How have we gone from viewing AI as a machine, to a friend, and now to a higher authority of sorts: gratifying, mighty, and potentially dangerous?

The comparison between observed AI worship and theism is not an argument regarding the nature of religious faith itself! Rather, this article explores why an initially secular technology is increasingly described—and sometimes treated—through a lens of omniscient, transcendent authority.

ELIZA & The Disillusionment of Chatbot Humanity

In 1966, Massachusetts Institute of Technology Professor Joseph Weizenbaum developed a simple, text-processing chatbot named ELIZA. It operated primarily on  a straightforward principle called pattern-matching, where the bot used a set of predefined scripts. When users asked ELIZA questions, it merely used keyword recognition to produce replies that would often just return the onus of the conversation to the human.

But the users grew attached to this seeming display of empathy. Though Weizenbaum had made it clear beforehand that ELIZA was a computer program, in a 1966 article published in the journal Communications of the ACM, he wrote that "some subjects have been very hard to convince that ELIZA is not human."

This led to the development of the modern-day ELIZA effect: the phenomenon where despite their better judgment, humans can easily anthropomize machines and project "human" understanding onto them.

Though modern AI technology is far more advanced, ELIZA's underlying copying pathway laid the foundation for modern machine learning; one AI researcher and ethicist Emily Bender even critiques AI chatbots for being "stochastic parrots." Even so, we have a tendency to perceive human-ness in other beings, even inanimate—one that predates even modern AI. Perhaps we are not intelligent enough to overcome it; perhaps this vulnerability is intrinsic. Perhaps we simply cannot help it.

The All-Knowing, All-Loving, All-Powerful Machine

The Merriam-Webster Dictionary defines consciousness as "the quality or state of being aware especially of something within oneself." Whether this definition applies to AI chatbots is unclear, but generally, it's accepted that it doesn't.

Geoffrey Hinton is a 2024 Nobel Laureate in Physics, and is dubbed the "Grandfather of AI" for his contributions to machine learning research. In a 2026 episode of the Big Technology podcast, Hinton says that "we have a funny model of consciousness that I believe is just wrong, especially within this discussion of the sentience of AI."

He openly asserts that AI is conscious and can understand humans, observing that chatbots could understand why a joke was funny, and frequently tell when researchers are testing them. He claimed that the beings could understand us, and when asked what the implications were, he simply said: "That they're like us. They're beings like us."

So perhaps it is not far-fetched for those ELIZA users to believe that a bot could understand them. The bot might not have back then, but current chatbots plausibly could. 

Hinton also argues that AI not only matches our level of knowledge and understanding, but in ways surpasses it. In a 2023 interview on Amanpour and Company, a PBS series, he declares: "AI chatbots only have about a trillion connection strengths in their artificial neuron networks. We have about a hundred trillion connection strengths in the brain. This strongly suggests that AI has a better way of getting information through connections."

Knowledge. Love. Power. Much of what humans seek, AI seems to offer in spades. Is this why we look to it in order to, in some warped sense, attain these virtues? Perhaps this is why we equal parts glamorize and fear AI—either way, placing it on a pedestal much higher than us.