When we imagine the existential threats posed by artificial intelligence, our minds often jump to cinematic tropes: killer robots marching through the streets or a rogue supercomputer deciding to eradicate humanity. But according to historian and philosopher Yuval Noah Harari, the true danger is far more insidious, yet equally catastrophic.

The greatest threat of the AI revolution, Harari warns, is not extinction by laser fire, but a slow descent into the “collective dementia of humankind.”

Just as a patient with dementia might be kept alive by machines while losing the ability to comprehend their surroundings, humanity risks surrendering its understanding of the world to algorithms. If we do not actively manage this transition, we may soon find ourselves living in a society where the systems governing our lives are entirely alien to us.

From Tools to Autonomous Agents

To understand the trajectory of AI, Harari argues, we must first recognize a fundamental shift in human invention. Throughout history, every piece of technology humanity has created—from the stone knife to nuclear energy—has been a tool.

A knife does not decide whether to chop a salad or commit a murder; the human holding it does. Nuclear technology does not decide whether to power a city or destroy it; politicians and scientists make that call. Tools require human intent to function.

AI is fundamentally different. It is not a tool; it is an agent.

For the first time in history, we have created something capable of making its own decisions and generating new ideas independent of its creators’ intentions. Billions of these agents are already embedded in our phones, banks, and governments. AI does not think like a human, and it never will. Harari likens it to airplanes: an airplane doesn’t fly by flapping its wings like a bird, but it achieves flight in a completely different, often superior, manner. AI is an “alien intelligence” operating in ways we are increasingly unable to predict.

Bureaucratic Natives and the “Cows of the Future”

If AI won’t conquer us with physical force, how will it take over? By mastering the very systems that gave humans dominion over the Earth: bureaucracies.

Humans dominate the planet because we can cooperate in massive numbers through abstract bureaucratic systems like laws, religions, and financial markets. These concepts are invisible to other animals. As Harari points out, you cannot explain the bond market to a cow, even though the agricultural financial system dictates that cow’s entire life.

For humans, bureaucracy is a learned, often clumsy skill. We evolved in the forest, not the office. But AIs are “bureaucratic natives.” They were born in the digital office. Because no human can retain every transaction in a global bank, AI will inevitably take over the financial and legal systems, optimizing them into networks so complex that no human brain can comprehend them.

Within a decade, Harari suggests, AI could create a financial system so sophisticated that we will understand it no better than a cow understands the stock exchange today. We will become the cows of the future.

The Invasion of “AI Immigrants”

This loss of control extends to the geopolitical stage. Currently, the AI race is dominated by two superpowers: the United States and China.

Harari views the global deployment of AI software as a wave of “AI immigrants”—code crossing borders at the speed of light, requiring no visas, and taking human jobs across all sectors, from taxi drivers to doctors and lawyers. More alarmingly, these digital immigrants hold zero political loyalty to the countries they operate in.

If smaller nations or unions—like the Czech Republic or even the broader European Union—run their critical infrastructure (hospitals, airports, militaries) on AI models controlled by foreign corporations, they effectively surrender their sovereignty. A switch flipped in Washington or Beijing could shut down an entire nation’s infrastructure, creating a form of imperial power unprecedented in human history.

Hacking Our Evolutionary Code

While the future holds bureaucratic dominance, Harari notes that AI algorithms have already caused massive societal damage by taking over our primary information ecosystem: social media.

In the pursuit of maximizing “user engagement” (and thereby profits), social media algorithms ran experiments on billions of human guinea pigs. They quickly discovered a flaw in human psychology. While the truth is often complex, costly, and boring, outrage is highly engaging.

Our brains evolved on the African savanna to immediately focus on threats. If someone yells “Snake!”, you pay attention. Today, social media algorithms flood our feeds with endless digital “snakes”—fear, greed, and outrage—hacking our evolutionary programming. This diet of mental junk food is directly responsible for the erosion of trust in modern society, making it incredibly difficult for humanity to agree on basic facts or cooperate effectively.

The Myth of Inevitability

Despite these daunting challenges, Harari rejects the narrative spun by tech billionaires like Elon Musk, who frame the AI revolution as an unstoppable train we are merely passengers on. This technological determinism is dangerous because it paralyzes political action.

There is still time for humanity to adapt, provided we prioritize safety over speed. Harari calls for urgent regulations, such as holding corporations legally liable for the actions of their AI agents, banning AIs from masquerading as humans, and requiring “information diet” labels on social media content to warn users of high levels of outrage or fear.

The future is not yet written. But if we fail to hit the brakes and assert our political agency, we risk trading our hard-won civilization for a state of collective dementia, wandering blindly through a world we built but no longer understand.