Harari, a professor of history at the Hebrew University of Jerusalem and author of Sapiens: A Brief History of Humankind, said AI could create systems so sophisticated that even the people living under them would no longer understand how they worked.
"I think the much bigger danger is something like collective dementia that we just lose the ability to understand the world," Harari said in an interview with Veda 24.
He compared the possibility to dementia in individuals who, despite remaining physically alive, gradually lose their understanding of their surroundings and control over their lives.
"Like people who are very strong and in control of their lives, and when they suffer from dementia as they get older they just lose the ability to understand what is happening," he explained. "They don't understand what is happening and I fear the collective dementia of humankind."
Why Harari believes AI could take over bureaucracies
Explaining his concerns, Harari said humanity's dominance over the planet did not come from individuals being physically stronger or more intelligent than other animals. Instead, humans developed complex systems that allowed millions of people to cooperate.
These include banks, financial markets, courts, legal institutions, governments and religious organisations. "The way that humans took over the planet is by building bureaucratic systems," he said.
According to Harari, these institutions became the foundation of human power because they allowed large populations to organise and work together in ways other species could not.
But AI, he argued, could prove far better suited to operating such systems than humans. "The AIs are taking over the financial system. They are taking over all the bureaucracies because they are bureaucratic natives," Harari said.
He explained that bureaucracy often feels unnatural to humans because human beings evolved long before modern institutions existed. "We were born in the forest, not in the office. So it's alien to us. For AI, bureaucracy is home. They're born in the office. They are bureaucratic natives."
The professor gave the example of banking, where a human employee cannot possibly remember every transaction conducted by a financial institution, while an AI system could process and retain enormous amounts of information.
He warned that this advantage could eventually allow AI to design and manage financial and legal systems that exceed human understanding.
"It's not that AIs will create robots that will come in the street and shoot us. The AIs are taking over the financial system," he said. "They will take over the bureaucracies like the financial system, create much more sophisticated financial systems and legal systems that we can't understand."
'We are the cows of the future'
To explain what losing control might look like, Harari compared humanity's possible relationship with AI to the relationship between humans and cows. A cow, he explained, can recognise humans but cannot understand the financial arrangements through which people buy, sell and trade cattle.
Even though those systems determine much of the animal's life, concepts such as bank transfers, stock exchanges and bond markets remain completely beyond its understanding.
Harari warned that humans could soon find themselves in a similar position. "So just as cows can't understand the financial system that controls their lives today, we will not be able to understand the AI financial system that will control our lives in 10 or 20 years," he said.
When the interviewer suggested that humans could become "the cows of the future", Harari agreed. "We are the cows of the future," he replied.
His argument was that the most consequential form of AI control might not involve physical force. Instead, it could emerge through financial and administrative systems that humans increasingly depend on but cannot comprehend.
Harari agrees with Elon Musk's prediction about money
Harari also referred to a prediction by Tesla CEO Elon Musk that money could cease to exist within the next decade. He said he agreed with Musk's assessment. "I disagree with some things he says and does, but I agree with that," Harari said.
He suggested that AI could eventually replace the financial arrangements humans have developed over thousands of years with systems that people may no longer understand. "This will disappear in a few years perhaps as AIs create new financial systems that we can't understand in the same way that the cows can't understand our financial system," he said.
When asked how people might buy everyday essentials such as bread, Harari acknowledged that the future remains highly uncertain. "Nobody knows. We have no idea," he said.
He explained that while some AI-related dangers are relatively easy to imagine, the most significant changes may involve systems and institutions that humans cannot currently predict.
For example, Harari said it was possible to imagine terrorists using AI to develop a dangerous new virus with the infectiousness of COVID and the lethality of AIDS. Such threats, he argued, should be taken seriously and addressed through precautions.