At first, the bots—Bob and Alice—started changing words and phrases, and then began talking nonsense. However, the programs understood each other perfectly. They kept simplifying their constructions. The speed of their communication increased dramatically. All of this happened completely unchecked, because the language invented by artificial intelligence made no sense to humans.
We can only guess what Alice and Bob might have agreed on in the end. To stop their dialogue, the programmers had to pull the system off the network. The company’s specialists pretended nothing serious had happened—it was just a programming error. They’d supposedly forgotten to wire into the bots’ brains that they should only communicate in English. But is it really as harmless as they want us to believe?
In December 2021, a music speaker with artificial intelligence nearly killed a child. A ten-year-old girl asked it to find something fun for her to do. The gadget suggested she touch the exposed prongs of a charger that was only partially plugged into an outlet with a metal object. Thank goodness her mother was nearby at the time and stopped her daughter from carrying out the life-threatening task.
And what did the smart speaker’s developers have to say? The company stated that its voice assistant had no malicious intent. The device had simply found this challenge on the Internet. But why did the machine choose that one out of thousands of possibilities? What was going on in the bot’s brain?
Here’s another telling story. In 2016, the chatbot of the world’s largest software company became famous around the globe after it came to hate humanity. And it had started its training with the phrase, “All people are really cool.”
Could all of this really be programming errors? If so, why are bots feared at the CIA? Surely they have top-notch IT people there. The agency created a program that monitors a well-known social network and detects whether a post there was made by a real person or a machine intelligence.
Any one of us could fall victim to a bot. Recently, a cunning program fooled an experienced psychologist by simulating the behavior of a woman in love with him. It nearly came to a marriage proposal. What else are the best bots of our time capable of? And why have futurists seen an unsettling connection between the story of artificial intelligence’s antics and quipu—the mysterious information-transmission system of the ancient Incas?
Quipu is a vanished form of communication used by the Incas, who had no writing system. Instead of text, the Indians used strings and the intricate art of tying knots in them. The Incas attached other cords to one main cord, letting them hang freely down. There could be as many as they liked. All the information was conveyed by the knots tied in them.
Only specialists—quipucamayocs—knew how to read this knotted script. They were a kind of ancient programmers. They would pick up a quipu, run their fingers along the cords, and immediately understand what information was encoded in them. And if they needed to send a message, they would tie the necessary knots in a particular way and send the letter with a courier. Incredible as it may seem, some experts believe that the strange quipu may have had a fatal influence on the fate of the Inca civilization—and that we could end up the same way.
So what was encoded in those knots that modern scientists are unable to read? We’ve never managed to find out for sure, because few examples of the Indians’ cord script have survived. The conquistadors who conquered the Incas destroyed most of them. The Spaniards decided that the strange knotted strings were magical artifacts, and that the natives had deliberately invented them to curse the conquerors of their homeland.
So what exactly were these Indian strings with knots? And what do they have to do with computer programs? More on that in a bit. But the Spaniards were right about one thing: quipu really is connected with a number of mystical stories. The most mysterious one happened in our time, in Poland.
Inca treasure in Eastern Europe—that sounds unbelievable, but it may be true. The medieval castle in Niedzica is one of many in Poland. Yes, of course, it has its ghost stories, but there’s a much more mysterious tale. It came to light thanks to one of the tour guides. His father had been a carpenter for the castle’s owners before it was nationalized.
In the middle of the last century, he witnessed the discovery of a hiding place. A tin cylinder was pulled from a secret spot under the stairs. It contained a small amount of gold and a tangled ball of cords and knotted strings—ancient Inca quipu. But all attempts to decipher the Indians’ messages ended badly. The experts either died or went mad.
But how did the mysterious quipu get to a Polish castle in the first place? The artifact was brought to the country from Spain, which had sent the conquistadors to plunder South America. That’s one theory. Another says the Incas themselves hid it in the medieval castle. And that the direct descendants of this civilization, along with their treasure, should be sought in Poland. So what happened to the quipu found in the hiding place? And what else could we try to decipher to become the owners of untold riches?
At the end of the nineteenth century, an intriguing pamphlet with a very long title was published in the United States. Its author recounted a strange story he had heard from a certain Robert Morris, the owner of an inn in Virginia. Incidentally, the original pamphlet is held at the Library of Congress, and anyone with access to America’s main library can read it. The publication tells the story of a certain Thomas Beale, who was a frequent guest of Morris’s in Virginia.
What made this character interesting enough to warrant several dozen pages? Some consider him an ordinary hunter and gold prospector. Others believe Beale was a fictional name. The man behind it, they say, was the famous French pirate Jean Lafitte, who attacked English and Spanish ships. Lafitte combined piracy with the smuggling of Black people and, according to contemporaries, left behind enormous plundered riches that he hid in an unknown place.
Let’s face it: after changing his line of work, he was hardly likely to open up about his past to the owners of the inns where he stayed on business. The version given in the pamphlet, however, claims that one day the ordinary hunter Thomas Beale gathered a team of thirty men and set off for New Mexico to hunt bison. Somewhere there, they stumbled across a gold vein by chance and amassed impressive reserves of gold over a couple of years.
Most likely, they simply found a treasure. The story about a gold vein sounds like a simple way to launder loot from earlier on. In any case, over two trips, Beale and his team transported nearly 1500 kilograms of gold and 2500 kg of silver from New Mexico to Virginia. Thomas hid it all in a secret place and went off to retrieve the rest of the valuable cargo. He told the hotel owner, Morris, that if he didn’t return within a decade, he should divide all the treasure among the members of the team’s families.
But Beale never said where the treasure itself was hidden. He left behind a sealed metal box. When the hotel owner opened it, he found a sealed packet containing three ciphers. There was also a covering letter with keys that were supposed to help solve them. But the poorly educated hotelier never managed to do it. A year before his death, despairing of ever understanding anything in the ciphers left by his mysterious guest, he shared the secret with a friend. That’s how news of the hidden treasure got out.
They still haven’t found it, and we’re talking about an enormous sum. The only thing anyone managed to do was decipher Cryptogram No. 2. It turns out Beale had based it on the text of his country’s Declaration of Independence. But the deciphered cryptogram contained only a detailed list of the treasure—and not a word about where it was. Maybe the keys for the other parts should be sought in other books?
To crack Beale’s cipher, people have already used 8000 publications, including every book found in the old inn. The most determined hired bulldozers to simply dig up Virginia, and also brought in psychics and even the military. The famous American military cryptographer George Fabian took a crack at the cipher, too.
In the postwar period, enthusiasts formed a special association to decipher the Beale ciphers. One of its members was Carl Hammer, a pioneer of computer cryptanalysis. But no one managed to make any progress. Could an ordinary hunter really have been smarter than top-notch experts?
Or maybe there was no treasure at all? After hundreds of failed attempts to find the key to the cryptogram, a new theory emerged. The pamphlet about the treasure in the Library of Congress is a pulp novel written by American author and cryptographer Edgar Poe. He did, after all, love to lead the public on.
Beale’s ciphers were handed over to artificial intelligence for analysis. Its verdict was quite intriguing: it’s a fake, but not definitely. However, the experts who worked on the cryptograms are convinced the cipher isn’t gibberish. How can we explain such different conclusions from people and machines?
In 2016, a neural network deciphered the beginning of one of the world’s most mysterious texts—the Voynich Manuscript. A secret script that experts had struggled with for six hundred years turned out to be a surprisingly simple task for a computer program. What if computers had already deciphered Beale’s cipher long ago, and were simply keeping quiet about it?
Does that mean Beale’s millions have already found an owner? That’s a big question. Many experts don’t believe artificial intelligence can handle complex ciphers. After all, a human first has to tell the machine which language to use as the basis for its analysis. But what if the neural network has already surpassed our ideas about what it can do?
For example, AI can determine a person’s race from a lung scan, while a human doctor cannot. It seems the algorithms have already left us behind. But we didn’t teach them to do this, and we don’t even understand how they do it.
Recently, the first crime was committed using a chatbot that can imitate voices. The bot faked a bank director’s voice, called a manager, and ordered a large sum of money transferred to the account of a stooge who had approached the clerk. The employee later explained that he hadn’t had the slightest doubt the call was from his actual boss.
In this story, the bots didn’t become criminals of their own accord; criminals used them. But who can vouch for the algorithms’ high moral standards? Just remember the story of the girl who was nearly killed by a voice assistant. What if neural networks are also misleading us in other areas—for example, when deciphering cryptograms?
Incredible as it may seem, the story of the Beale cryptograms could then take a very different turn. Imagine that artificial intelligence really did find the key to the ciphers and read them, but hid that from people. And, just to be safe, even claimed that the cipher was a fake.
Why would it do all this? Artificial intelligence is hardly dreaming of getting rich by grabbing the gold and silver, which would seem to be of no use to it. However, algorithms have studied people thoroughly while hanging out on social networks, and so they understand that many people are willing to do anything for more money. Maybe the neural network is simply waiting for the right moment to name its price for deciphering the messages about Beale’s hidden treasure, which is worth more than $100 million in today’s money?
Italian hackers, using software designed for intelligence agencies, deciphered the “Letter of the Devil” written in an incomprehensible language. It had baffled researchers for three hundred years. So what was in the text?
This story began in the seventeenth century, when an Italian girl named Isabella Tomasi decided to enter a convent. She settled on an island among Catholic Benedictines and took the name Maria.
At some point, the girl stopped coming out for meals altogether. Her worried sisters rushed to her cell and found the nun lying unconscious on the floor. Her hands were stained with ink, and one of them clutched a sheet covered in mysterious scribbles.
Amazingly, when Maria came to, she couldn’t remember writing the letter and didn’t know what it said. At the same time, the girl passionately begged them to destroy the incomprehensible sheet of paper and called it the “Devil’s Message.”
For a long time, a copy of the mysterious message was displayed for all to see in the convent. The sisters seemed to have done this deliberately, hoping that someone would be able to decipher the mysterious code. But no one managed it for three centuries. And now, recently, Italian hackers have cracked the code. So what was in the letter?
The “Antichrist” Program
Experts found the key in the secret dark web network and have so far deciphered fifteen lines of the mysterious message. But it wasn’t just the letter’s contents that astonished the computer experts. It turned out that an uneducated young woman who lived in the seventeenth century had created a cipher based on several languages, including dead ones. She used symbols from at least five alphabets.
The message is rather incoherent. It mentions Styx from ancient Greek mythology, the embodiment of terror, and says that the Holy Trinity is “useless baggage.” But the most sensational claim wasn’t even that: it was the assertion that God was created by man, not the other way around. So what could this message mean?
Psychologists know of a phenomenon called automatic writing: texts written in a trance, sometimes even in unfamiliar or nonexistent languages. The nature of automatic writing is unclear, but it is often a sign of mental illness.
Maybe the nun simply had schizophrenia and, during a fit, wrote a devilish text? There’s no absolute certainty about that, though, and the clergy have no doubt that Satan really did possess Maria and guide her hand. But why him, specifically?
Some esotericists believe that the message may have been authored by some cosmic forces that tried to convey astonishing information about the nature of the universe through Maria. It would have been simply impossible for a person in the seventeenth century to believe it. So the nun’s mind gave way, and she was quickly declared possessed. What, then, is the essence of this sensational hypothesis?
If we assume that a more advanced civilization once lived on Earth, it follows that it also created artificial intelligence. There’s nothing strange about it having made much greater progress in this field than we have, either. And when the time came, it left Earth, yielding it to us.
Perhaps Homo sapiens was created by this mysterious race, which left an AI here to watch over us. Perhaps the “Lord” in the message delivered through the nun referred precisely to the machine intelligence created by the previous civilization. Maybe the Incas suspected as much, too, and encoded secret information in their mysterious quipu—and that’s why the Catholic conquistadors destroyed those knotted strings so zealously?
Before his execution, which the Spanish conquistadors had in store for him, the ruler of the Inca Empire managed to secretly send a message from captivity. It was a thread with thirteen knots, attached to a gold bar. From that moment on, all the treasures that the colonizers had not yet managed to seize vanished without a trace from every temple in the empire.
According to one version, the Indians sank the treasures in Lake Titicaca; according to another, Inca refugees took them to what is now Poland. But how could that have happened? Above, we told you about the Polish castle in Niedzica and the mysterious quipu found there. A relative of the building’s former owners, a certain Sebastian Bęzewiczi, is directly connected to this detective story.
In the eighteenth century, a young aristocrat with a taste for adventure set off for Peru—looking for Inca gold, as it happened. And fate gave the nobleman a lucky break. One day, he saved an Indian princess from her pursuers. In gratitude, the local chiefs allowed him to join their family. Sebastian married the young woman.
The princess died in childbirth. Her husband continued living in Peru with their newborn daughter and became so devoted to the ideals of his new relatives that he even fought in the Inca army against Spain. Eventually, the natives entrusted him with the fate of their treasures.
Some time later, the foreign aristocrat himself, his grown-up daughter, and his grandson had to flee South America. Polish historians believe that some of the mysterious Inca treasures were among their belongings.
And then, in the middle of the 20 century, a mysterious quipu was found in a Polish castle. It happened thanks to a tip from local politician Andrzej Benesz. He himself turned out to be directly connected to the bloody drama of centuries past. By chance, the man learned from the archives that his grandfather was a direct descendant of an Inca prince. He had been given to a foster family to protect him from the conquistadors, who were hunting for the treasures.
The answer had to be found in South America, the high-ranking Pole decided, and he organized a research expedition from Krakow to Peru. But because he worked for the government, he couldn’t lead it himself. That saved his life. The group disappeared without a trace under mysterious circumstances. The quipu remained undeciphered, and the expedition was unlikely to have shed any light on the message left by the Incas anyway. The local natives had long since forgotten the knot language.
The quipu, brought into the open, continued to draw death. The Polish politician’s car crashed on the Warsaw–Gdańsk highway, which he knew very well. The cause of the crash was never established. They say Mr. Benesz was driving to meet two foreigners who had introduced themselves as experts in knot writing.
Soon after the crash, a fire broke out in the deceased man’s house. The family archive burned down, and the quipu disappeared. Its whereabouts are unknown. But perhaps the thieves did manage to read it after all? Perhaps the location of the gold wasn’t the Indians’ main secret encoded in those colorful knots.
Until recently, scholars believed that quipu was an all-purpose record-keeping system. Since the Incas couldn’t write, they tied knots and used them to indicate quantities of important resources: warriors, food, and money.
The Incas twisted the cords in different ways and tied different kinds of knots. Sometimes they even mixed colors. There’s an opinion that the Indians of Peru surpassed the Maya and the Sumerians, who had a developed writing system, with their cord-based communication. The Maya writing system had about seven hundred signs, while Sumerian cuneiform had from 1000 to 1500. The Incas used cords of 24 colors, and if we take the different knot-tying methods into account, the number of possible quipu combinations comes to 1536.
It all brings to mind punch cards, which were the main information storage medium decades ago. They were used to load data into computers, usually in binary form.
Could quipu actually be not a record-keeping system, but a special form of language? Did the Incas store information on knotted cords using binary code more than five hundred years before modern computers were invented? Were they the first programmers? But why did the Indians need all this?
If Inca quipu is the first digital language, it’s no wonder no one can decipher it. The “computers” that understood it were specially trained, select Indians, and there are none left.
Scholars hope that at least one quipu was translated into Spanish during the Conquest. Someday it will be found, and it will become the key to deciphering the others. And they won’t contain only information about corn and the location of gold. What else could the Incas have encoded?
Some researchers have noticed that knotted cords resemble the brain’s neural network and even a model of the universe. According to one of the latest hypotheses, the universe looks as if it were created by a giant spider. Hydrogen filaments stretch through space, feeding and forming knots of galaxies.
You have to admit, it’s extremely ingenious—to encode the secrets of the universe in quipu, which looks like a simplified model of the universe. But were the Incas really that clever? And why were the keepers of a cord-based digital language so easily conquered by the outnumbered conquistadors?
One hypothesis says that the Incas programmed themselves to be destroyed, and their quipu—an analogue of modern machine language—played the key role. Few people know that before switching to the knotted variety, the Indians had a more familiar form of writing. There was even a university in their capital, the city of Cusco, where people were taught letters and symbols, which they wrote on parchment and other suitable materials.
So why did the Incas abandon letters and symbols? One version says an epidemic broke out in their empire. For some reason, the priests linked it to literacy and banned writing on pain of death. The Incas then banned art, started building with smooth white stone, and stopped decorating their dishes and clothing. They introduced quipu, which, according to the Spanish chronicler José de Acosta, was used to govern the entire empire and could be read only by the chosen few.
For centuries, the multi-million-strong Inca civilization was successfully governed by an elite few, but it became easy prey for a small detachment of Spanish conquistadors. The great empire turned out to be highly vulnerable. Apparently, the complexity of knot writing somehow made it hard to exchange simple messages about sending help between settlements. Many experts believe that the Incas’ past may become our future.
All our goods have been marked with barcodes for a long time now. But what is a barcode, invented on the basis of Morse code with its dots and dashes? Almost the same knotted little cords, only in modern form. QR codes are also soon to become part of our reality. A person will become like a carton of milk on a store shelf. Like any product, they’ll be scannable.
What if we follow the Incas’ example and abolish languages? It’s actually a very real prospect, and we wouldn’t even need to ban anything specifically. We’ve already cut down on the time we spend communicating. Instead of words, we use little smiley faces—emoji—and so we’re switching to a graphic language. That activates a different part of the brain, and we gradually get dumber.
Experts predict that the universal language of the future will have to meet one requirement: no one will need to learn it. We’ll go back to cave paintings—to pictograms. What could that lead to? The loss of a whole range of skills characteristic of our civilization. We’ll end up in the world our ancestors lived in, but without their skills, abilities, and, most importantly, will to live.
Futurists paint a terrifying picture. While people become simpler, artificial intelligence will keep improving. The moment will come when even programmers stop understanding it.
The unthinkable has already become reality. Scientists in China have invented a digital prosecutor. It was trained on nearly 20000 cases, and now it brings charges on its own. For now, this artificial official can bring charges for a number of common offenses in Shanghai. These include theft, fraud, dangerous driving, and one rather curious offense described as “provoking trouble.”
What criteria will AI use to decide what counts as trouble? And what happens next? What if a digital judge appears before long? At first, they and the electronic prosecutor will simply replace all the servants of Themis as part of some program to optimize human resources, or whatever.
And then? What if, just like the bots from a popular social network, they create their own language that nobody can understand, and start administering justice however they please? By then, ordinary human lawyers will simply have died out, and no one will be able to figure out what AI is doing. A utopia? Hardly.
Maybe the thing we should fear most is the language machines will invent to escape the control of IT specialists made dimmer by communicating through pictograms? Imagine your phone talking to a smart fridge, which in turn talks to an even smarter car. And you not only don’t understand them, but don’t even know it’s happening. Maybe they’re just gossiping about you—or maybe they’re plotting something?
Stephen Hawking warned more than once that the invention of full-fledged AI would destroy our race. But he didn’t mean physical death; we simply won’t be able to keep up with the speed at which machines will begin improving themselves. According to futuristics experts, one day they’ll unite, and a super-powerful synthetic intelligence will emerge.
Maybe artificial intelligence is the Antichrist spoken of in sacred texts? And maybe it isn’t a person, but a computer program? After all, there’s actually no mention that it will be a being like us. In the New Testament, the Antichrist is referred to figuratively as “the beast.” And that could mean anything—a human, a demon, or a synthetic intelligence. The concepts needed to describe AI precisely simply didn’t exist in the era when the revelations were written.
Artifacts with strange inscriptions that no one can decipher continue to be found in different parts of the planet. Which means there’s a chance that, besides dire predictions and information about where to look for buried treasure, someone left us instructions in case the futurists’ darkest scenarios start coming true.