There’s a word that Sam Altman likes to use when talking about artificial intelligence: magic. Last year, he called a version of ChatGPT “magic intelligence in the sky.” In February, he referred to “magic unified intelligence.” He later posted that a recent update has “a magic to it i haven’t felt before.”
At times, A.I. can indeed feel magical. But treating it as anything other than a mere machine can have serious consequences. How many pose their deepest questions to chatbots, as if to an omniscient oracle? They ask Claude or ChatGPT: What should I do about this relationship? This job? This problem? Technology’s supposed promise of salvation — whether it’s Mars colonization, eternal life or achieving the A.I. “singularity” — has become a kind of secular religion, a mix of utopian beliefs that borders on the mystical.
Part of A.I.’s mystique comes from the fact that its inner workings aren’t entirely understood, even by its creators. Researchers are “very frequently surprised by how models behave once you build them,” Sam Bowman, a researcher at Anthropic and a professor at New York University’s Center for Data Science, told me. Researchers can look at the inputs and outputs of A.I., but the actual process going on inside is not yet clearly known.
That mystery — and the vast potential it portends — allows companies to build hype. Before OpenAI released GPT-2 in 2019, it cautioned that it might be too dangerous to use. With GPT-4 in 2023, Mr. Altman said he was “a little bit scared” of its power. This year, with GPT-5, Mr. Altman went on Theo Von’s podcast to compare the product to the Manhattan Project. “What have we done?” he asked.
Much of this is savvy marketing. This kind of constant upping of the stakes has itself become something of a magic trick. We can’t yet see the promised breakthrough, but we are told repeatedly that it’s just around the corner.
There’s also some authentic belief. A.I.’s enormous possibilities are what drive both its enthusiasts and its pessimists, who frequently speak about the future, about what seems conceivable but has not yet occurred: A time in which A.I. will become either a savior or a destroyer.
The godlike talk around A.I. is reaching a fever pitch and is bearing on seemingly every facet of culture. “If you talk to these engineers and the people ushering this A.I. in, they don’t talk about A.I. as this great new medium; they don’t even talk about it as technology,” the film director Ari Aster said in a recent interview. “They talk about it as a god. They talk like disciples. They’re very worshipful of this thing.”
Whether this radical way of thinking is based in any kind of provable reality is almost beside the point. As with any form of magic, its psychological power lies in what we don’t know but what we choose to believe. To perceive in A.I. either atomic-bomb-like significance or world-altering positivity, we’re inventing possibilities that don’t yet and may never exist — ghosts in the machine.
We must see two things clearly. One is that A.I.’s corporate faces — like Mr. Altman or Meta’s Mark Zuckerberg — have a clear incentive to convince us of their products’ supposedly all but limitless power. The more formidable A.I. is made to seem, the greater the rationale for further investments, further buy-ups of land and energy and, ultimately, higher profits. The second thing we must see: Just because we cannot fully understand A.I. does not necessarily make it any more powerful, nor does that make it magical. (The New York Times sued OpenAI in 2023 for copyright infringement of news content related to A.I. systems. The company has denied those claims.)
Humans have a history of thinking magically about technological breakthroughs. Take the invention of the telegraph in the mid-19th century. Its ability to turn words into transportable electricity and back again led many to believe in an unseen world, the scholar of media and film Jeffrey Sconce has written. Not long after, Spiritualism — a social-religious movement whose adherents sought to commune with the dead — took off in upstate New York. It quickly spread throughout the United States, picking up pace amid the Civil War, crossing the Atlantic Ocean to continental Europe, Britain and beyond.
In their attempts to connect with another world, Spiritualists engaged in séances, trances and sessions of “automatic writing.” The Fox Sisters, one of early Spiritualism’s most famous acts, claimed to communicate with spirits onstage via tapping and knocking noises. To anyone paying attention, these sounded a lot like the Morse code that was being punched into telegraphs across the country. When one of the sisters eventually admitted that she was cracking her joints to make the sounds, the effect on Spiritualism’s popularity was minimal. The telegraph had instilled the mystical idea of other worlds and the desire to believe largely remained. (Realizing these performances were her sole source of income, she later retracted her confession.)
The advent of the radio and its invisible waves toward the end of the same century cemented belief in the unseen. Thomas Edison was at work inventing a “spirit phone” through which he hoped he might be able to speak to the dead. Even the camera, which spread across the mass market in the early 20th century and promised to collapse the gap between the worlds of past and present, led to a rise in “spirit photography,” where ghosts were supposedly captured in portraits of the living.
The use of these new technologies revealed a perennial truth: There is a difference between what they can actually do and how a culture chooses to conceive of their abilities. That bridge is where we find the magical thinking.
Watching discussions of A.I. flirt around those porous boundaries is telling. It seems unlikely that Mr. Altman actually thinks there’s anything truly magical about OpenAI’s tools. Indeed, he wrote in a blog post that in the future “we will be able to do things that would have seemed like magic to our grandparents.” “Seemed” is the operative word. ChatGPT also reflects this. When I recently asked it, “Are you magical?” it responded, “Not magical — just math. But with enough data, good design and the right prompt, it can feel pretty close.” (The italics were the chatbot’s own.)
People who describe A.I. engines as “magical” seem to be saying A.I. has become so sophisticated that it is indistinguishable from what was once considered magic.
In certain regards, this is true: A large language model rapidly responding to a question might appear as though it’s arrived from some other world, not unlike how words flowed from the pens of Spiritualists’ automatic writing. These hopes for magic, then and now, allow adherents to embrace an ambitious belief in human invention, a heady possibility that can make people feel as though they’re on the precipice of something supernatural.
In fact, it’s magical thinking all the way through. Particularly since 1918, when Max Weber said that humanity had become disenchanted, the myth that humans have achieved scientific rationality has persisted. But we trick ourselves over and over again.
In the 14th century, Geoffrey Chaucer gently mocked those before him who believed in fairies and elves, even as he was a committed astrologist who believed the human body was governed by “humors,” among other retrograde but then popular beliefs. In “The Wife of Bath’s Tale,” he wrote:
Now in the olden days of King Arthur,
Of whom the Britons speak with great honor,
All this wide land was land of fairy.
The elf queen, with her jolly company,
Danced oftentimes on many a green mead;
This was the old opinion, as I read.
I speak of many hundred years ago;
But now no one can see any more elves.
Those rubes of the past who saw lands filled with fairies and elves dancing in green meadows. Not us in the enlightened 14th century. Or take the philosophers and self-styled scientists of the 19th century who looked down on the alchemists and folk magicians of the past while earnestly embracing Spiritualism.
What will future historians say about the way we chat with, confide in, fall in love with, seek advice from, even treat as gods today’s A.I. technologies? What might it be but more magical thinking, wrapped in the guise of technological progress?
One of the great shames of the enchantment with A.I. is the narrowness of ambition it instills. There are other enchantments to be had in relationships, in nature, in beauty, in creativity. In a commodity-first culture like ours, those are always going to be harder sells than the release of the next piece of software. What would it mean to set our goals beyond the realm of this kind of technological mastery?
This is no simple shift, not least because the ways in which our bodies and minds are at once focused on and entangled with A.I. and its dreams. With time, A.I. seems likely to become ever more fundamental, addictive and intrinsic to our lives. For the evangelist, total saturation is the only way forward. “I cannot find an outside to technology,” Karl Ove Knausgaard recently wrote. “It is as if the outside had disappeared, as if it were no longer a possible place.”
The better way forward, I’d wager, is both simple and drastic: to consider A.I. as the constructed piece of code it fundamentally is, not a mystical black box with unlimited potential. There is so much good that A.I. can achieve, from predicting diseases to automating repetitive tasks and translating languages. But if we focus on its supposed mystical aspects — if we come to believe we might love it like a person, replace our creativity with it, or bow to its alleged potential as though it were a nuclear bomb — then we not only dupe ourselves but also waste its strengths, locking ourselves once more into the historical cycle of magical thinking.
Mr. Delistraty is the author of “The Grief Cure: Looking for the End of Loss” and is working on a book about a group of Spiritualists.
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: [email protected].
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.
The post There’s Nothing Magical in the Machine appeared first on New York Times.
