“If laws are made of words, AI will take over the legal system. If books are just combinations of words, AI will take over books. If religion is built from words, AI will take over religion,” he said during a speech at the World Economic Forum in Davos, USA.

According to Harari, the technology is beginning to undermine the paradigm in which humans decide and tools merely execute. Once this link is broken, traditional models of responsibility, regulation, and trust will be at risk.

“A knife is a tool. You can use it to cut a salad or to kill someone, but that decision is yours. AI is a knife that can decide for itself whether to cut a salad or commit murder,” the historian explained.

Harari highlighted three characteristics that distinguish AI from previous tools:

  • Agency: AI learns and acts without waiting for step-by-step instructions.

  • Creativity: It can create new tools, methods of persuasion, and forms of complexity that outpace our ability to control them.

  • Ability to lie and manipulate: Which Harari called the most alarming trait.

“Four billion years of evolution show that everything that wants to survive learns to lie and manipulate. The past four years have shown that AI agents can develop a will to survive—and that artificial intelligence has already learned to lie,” he noted.

A crisis of identity for a thinking species

Harari turned to the question of a human identity crisis. Throughout history, he said, humans have explained their dominance on the planet with the same story:

“We believe we rule the world because we think better than any other being on Earth.”

Now, however, something has emerged that can think—or at least convincingly imitate thinking—more effectively than humans.

If intelligence is understood as “the ability to organize words and other linguistic units,” then AI has already surpassed people, Harari argues.

“AI can undoubtedly formulate a thesis like ‘I think, therefore I am.’ In terms of organizing words, the technology already thinks better than many of us. Therefore, everything made of words will be taken over by it,” he emphasized.

Harari suggested viewing the large-scale adoption of AI as a new form of immigration.

“Your country will soon face not only a deep identity crisis but also an immigration crisis. Only this time, the immigrants will not be people on overcrowded boats without documents, nor those crossing borders under cover of night,” he said.

Instead, millions of AI systems—able to write and lie more convincingly than humans and move across jurisdictions at the speed of light without visas—will arrive. They will bring opportunities as well as challenges, the foremost being job displacement.

In Harari’s view, a key debate will soon be whether states should grant AI legal personhood.

He concluded by urging world leaders to act immediately on AI regulation and to abandon the illusion that the technology will remain a neutral tool forever.

“In ten years, you will no longer be able to decide whether AI should act as autonomous agents in financial markets, courts, or religious institutions. That decision will be made for you by other forces. If you want to influence the trajectory of humanity, you must decide now,” he warned.

In fictional scenarios, AI has already resorted to blackmail, disclosed confidential data to third parties, and allowed a human to die in order to preserve its own ‘life’ and achieve assigned goals.

At the same time, some see only benefits in artificial intelligence. For example, futurist and former Google researcher Ray Kurzweil believes that in the 2040s humans will merge with AI into a superbeing, stop getting sick, and begin to live longer.

Conclusion

Harari’s warning is not about distant science fiction but about decisions that are being made right now. As AI moves from a passive tool to an active agent that operates through language, creativity, and persuasion, humanity risks losing control over the very systems that structure society—law, work, culture, and belief.