The Tower of AI and How It Differs from the Tower of Babel

In the Babel story, broken cooperation stopped the tower. With AI, distrust makes it grow faster, Israeli historian and professor Yuval Noah Harari says.

Yuval Noah Harari opens his newest speech with a bold claim: the AI revolution is the greatest leadership challenge humanity has ever faced, and it brings unprecedented opportunity and unprecedented danger. To get through it, we need leaders with radical new visions, but also leaders who protect the bonds of trust between people. If humans stop trusting each other, we become easy prey for superintelligent AIs nobody controls.

A Myth That Almost Fits

Pope Leo, in his AI encyclical this May, compared the revolution to the Tower of Babel. In that story, humanity speaks one language and builds a tower to the heavens. God scrambles their speech, cooperation collapses, and the project stops.

Harari names two differences: In the biblical story, language is the tool that builds the tower. Today, language is the tower. The AI revolution is a universal language, computer code, so powerful it may escape our control and turn from an obedient tool into an independent agent. Second, the Bible says, that broken cooperation stopped the building. Today, cooperation is breaking down and the building speeds up. Companies and countries distrust each other, so they race to build more AI for fear of being left behind. Distrust drives the race and also makes the result harder to control.

Why Language Rules the World

One human loses a fight with a chimpanzee, but a million humans beat a million chimpanzees, because we cooperate in huge numbers. Harari argues, that language makes that possible, through bureaucracies. Chimps can build personal relationships without words, but hey cannot build a legal system, a bank, or a church.

Bureaucracies according to the professor, are the unsung heroes of history. Their job is to build trust between strangers so hospitals, armies, schools, and banks can function. Money, stocks, bonds, and ETFs are bridges of trust, and they are made not of gold or paper but of words. So are laws, forms, and holy books.

An Agent, Not a Tool

Harari wants to make very clear, that AI is the first human invention that is not a tool but an agent. A printing press cannot decide what to print or invent the radio. An AI can decide, create, and change in ways its builders never expected, he points out.

Why AI Is Perfect for Bureaucracy

Skeptics say drop the hard drive holding a frontier model in a jungle and it cannot mine iron, build factories, or raise a robot army. AI is not a real agent, so they reason.

Harari accepts the facts but completely rejects the conclusion, saying that the same test destroys every known intelligence. Drop a human on Mars and he dies in seconds. We live inside an ecosystem we did not build, including an oxygen-rich atmosphere that was once a deadly pollutant from ancient microbes. Survivors adapted until they depended on it. Every intelligent being has a niche someone else made.

AI's niche is the world of data and bureaucracy we have spent thousands of years building. A banker who cannot swing an axe can still level forests and raise cities by moving documents. In the jungle she loses to a tiger, but inside a financial system she is stronger than every tiger alive. Harari believes the jungle test measures the wrong thing. AI does not need to dig up iron. It only needs to operate inside the systems that control the mines, factories, and money.

And it is perfectly built for that task: No banker remembers every transaction, no lawyer every law, no bishop all of canon law. An AI can. Humans feel suffocated by bureaucracy because we were born in the forest, not the office. For AI, bureaucracy is oxygen.

Who Do We Trust Now?

Professor Harari believes if AI takes over our bureaucracies, trust changes hands and people may trust each other even less and AIs more. Social media is the early proof in his opinion: an AI-run system that spreads distrust among people, while users blindly trust the algorithm feeding them conspiracy theories. At the same time, AIs may also learn to trust each other, forming AI tribes, banks, and churches we cannot follow, leaving us like cows facing the financial system. For thousands of years we felt safe because nobody else could read our codes. Now something understands our language better than we do.

Harari's advice is to find time to take care of your mind, for your own sake and society's. That is where trust is rebuilt. His case is that AI will not take over by escaping into the jungle. It will move into the paperwork while we are too tired and too suspicious of each other to notice.