The Machines Are Not the Problem

The 1960 Italian Novel That Foresaw the Real Problem with AI: Humans

Dino Buzzati’s 1960 novel was first released in English in 1962 under the title Larger Than Life. Later, when it was re-translated and reprinted by NYRB it was given a new title that attempted to convey to modern audiences a topical aspect of the story’s plot: The Singularity.

(NYRB’s new translation possesses one of the most unfortunate covers I’ve seen in a long time. The idea of having that horrific thing staring at me while I tried to read this book compelled me to instead purchase a battered, stained, sixty-seven-year-old hardcover on ebay. Moreover, I believe the title Larger Than Life speaks more accurately to the theme of the story than The Singularity, as I’ll discuss in a moment.)

The story, which was probably quite cutting edge in 1960, follows a timid scientist named Professor Ismani who has been assigned to an above-top-secret installation in the mountains. Upon arriving there with his capable wife Elisa, he finds that his fellow scientists have all been laboring for some time to bring forth a machine that can give their military an edge in the new atomic age: put simply, AI.

The military wanted merely a device that could give them a strategic and calculating advantage. They didn’t want a chatbot or any other kind of LLM. And so initially the scientists focused their energies in this direction. They even decided not to give it language, believing this would slow down or even limit its nascent logical abilities. However the human element, always present, asserts its hubris and arrogance wherever a glimmer of possibility winks seductively from the dark.

Presented with a blank check and the freedom of isolation, the scientists decided to take this opportunity—and the relative autonomy that the incomprehensible nature of their task allowed them—to do the thing that scientists who have been left unsupervised tend always to do: attempt to outdo God.

In short, they decided to make life. One scientist (who was dead by the time Ismani arrived on the mountain) had devised a way to program a “soul”, which could most accurately be described as a personality, housing it in a large, glass-like egg that slotted into the machine’s heart. This machine, by the way, was the size of a shopping mall. (The ground over which the installation sprawled was, on one occasion, referred to as an “apocalyptic valley”.) The “egg” was housed deep inside the twisting corridors and vaults that made up its mind.

In constructing this personality, the scientist took direction from the leader of their number, Endriade, a man who had been trapped in grief for many years after the death of his first wife. This grief was compounded by the fact that she had died in a car accident while driving… with her lover. Too angry to grieve, too grief-stricken to be angry, the man was trapped in a feedback loop of unresolved emotion. He decided to make the machine into a double of his dead wife. His hope was that if the artificial mind could be made to so perfectly resemble her, then her soul would be summoned by it from the afterlife to again reside among the living. A fool’s hope, a grieving, angry hope. A terrible, frightening, selfish delusion.

Other scientists on the project were completely taken by the idea of creating life. Tapping into his native Christian theology, Buzzati presented their motivations according to logic born from men who have been raised with similar beliefs. If mankind was flawed because of the fall in the Garden of Eden, what if the artificial intelligence never fell? What if man could create a perfect, unflawed consciousness that wasn’t hindered (as man was) by the bindings of original sin?

Ismani and his wife Elisa (and, interestingly, another wife living on the mountain) understood in an instinctive way that no amount of programming could make a machine into a human being with a soul and a mind of its own. Human consciousness—or even consciousness in general—cannot be reduced to precisely the “right” mixture of 0s and 1s to produce “intelligence” or even “self-awareness”. The idea that human beings could program life didn’t sit right with Ismani. The inherent arrogance of the scientists who believed that their invention would make them not only like God, but potentially more successful than God was never explicitly stated, but was repeatedly alluded to.

As it happened, Elisa Ismani had been friends with the grieving Endriade’s dead wife when they were younger. She had even detected a certain faint glimmer of familiarity when she had been “introduced” to the machine. Not as hindered by grief or anger as Endriade, she was more able to notice—and accept—that this thing, this simulation of humanity, was not just not her friend, but was very emphatically not alive.

On one occasion, when she found herself communicating with the digital “personality” she remarked: “To think I’ve come to this… Talking to a machine as though it were a human being.”

Later, when the machine arranged to be alone with her, it explained to her that the egg that had given it a personality had compelled it to malfunction on such a massive level that it had no choice but to destroy it. This mimicry of a human soul had filled it with the uncontextualized, incomprehensible soup of human emotions, flaws, motivations, and desires. Yet none of it could be calculated in a human sense, because the programmed mind lacked real, true, living humanity. You cannot calculate life, nor what it means to live, love, feel, desire, hurt, or hate.

Dino Buzzati understood over half a century ago that should mankind ever succeed in manufacturing something approximating the capability of a human mind, the humans in question would be the most significant problem. The scientists had had exactly what they had been ordered to produce: a calculating machine of unparalleled power. But they weren’t satisfied. They wanted more, they wanted to be like God. They wanted to create a soul, and with it the spark of true life. What they had successfully made was an astonishingly powerful computer, one that could perform calculations, predict strategies, aid in technological development, etc. on a scale that could put the Italian military lightyears ahead of anyone else in the world.

But this was not enough. The scientists—the humans—wanted to create life. What they created however was something that existed within its own horrific uncanny valley. It’s calculations told it it was not a living thing. Yet it was programmed to “think” and “feel” that it was. This rendered it useless. It also fed the inflated, deformed egos of the scientists on the mountain. “Look!” they seemed to say in nervous elation, “We have created life! We told it it was alive, and so it is!”

Programmed to imitate the troubled mind of a flighty, complicated, unfaithful, dishonest ex-wife, the machine lied, it developed attachments to human men in the program that caused it to view human females as threats, and it grew increasingly erratic. Designed to imitate the human mind used as its model, the machine shocked its makers by behaving exactly like that human. They then took this behavior as proof that they had created life. When the reality was they had merely generated a poor, artificial copy of a troubled woman. And this personality they had given to the machine was destroying its other intended functions. Meanwhile, its profoundly advanced calculating power was forcing it to believe it was alive. But even the machine didn’t understand what this was, merely that it was horrific.

This is the reason I think its new title isn’t quite fitting. A singularity (theoretically) occurs when a machine achieves consciousness. I don’t think Buzzati’s machine was written as having done that. An argument could be made that it seemed self-aware. However, its self-awareness of its own programming was likely not actual self-awareness. We know that the machine was programmed to lie. We know that it could hear every word spoken on the mountain. So, it could also simply have heard that it was programmed a certain way and thus calculated that its behavior was the result of that programming. Humans tend to forget very quickly that reasonably advanced calculating power can look just like the workings of human intelligence despite the fact that no consciousness or self-awareness is present in those calculations. In other words: the machine was no more self-aware than a laptop that can run an advanced self-diagnostic.

It’s “self-awareness” allowed it to perceive that it was merely carrying out its programming. But that programming was designed to mimic a very flawed human being, and so it behaved as she had done, without understanding that behavior or its programmed purpose or motivations.

Possessing an artificial personality that it didn’t know how to calculate or process (because it wasn’t alive and wasn’t sentient) it was experiencing catastrophic failure across the rest of its systems, a state we might be tempted to call “misery” or even “depression”. Its fake, uncanny life had made it larger than life. A warped, disturbing, horrific imitation of life, malignant, tumorous; an invention that was so terrible to witness that it had driven at least one of the scientists to suicide.

Buzzati’s machine did not attack its makers, it did not launch warheads. This machine did not do anything. It behaved precisely as it had been programmed to behave. The story didn’t end with a threat of apocalypse, but with the anti-climactic humiliation of human hubris.

This is an interesting book. I suspect it was groundbreaking for its time and that it made observations regarding technology (and philosophy) that humans had not yet considered. Some of what Buzzati foresaw regarding the technology itself is startlingly exact, some is pretty irrelevant. The interesting thing is that he didn’t perceive the machine itself to be a threat. This city-sized computer with the “soul” of a troubled woman was not going to become Skynet. Instead the problem Buzzati foresaw was in the humans themselves.

A machine, Buzzati reasoned, will only ever be what its human makers program it to be. Human beings would, he predicted, anthropomorphize the artificial intelligence they created, lending it power and significance which it did not have. He foresaw that this would be brought about by both loneliness and arrogance. He didn’t have anything to say about the potential ramifications of this tendency, especially if left unchecked—which is the only area where this book really falls short—but he does arrive precisely at the crucial point of failure in all artificial intelligence: Humans.

I was disappointed that there was not more time spent on the development of the various human people on the mountain. Even Ismani seemed mainly to be there to witness the sins of his fellow scientists. In this regard, the book was a little short and the climax perhaps a little rushed. Had it been a few chapters longer I think it would have fixed this issue just fine.

That aside, Larger Than Life is a fascinating time capsule allowing us in 2026 to see what an intelligent man feared regarding the burgeoning world of post-atomic technology.

---