As Israeli historian, philosopher, and futurist Yuval Noah Harari explained in an interview, artificial intelligence (AI) has the potential to lead to an unprecedented concentration of power, which could prove significantly greater than that of previous technological revolutions. This was reported by Dengi.ua, citing an article in Mirror of the Week.
According to Harari, data and the infrastructure supporting AI systems are becoming the most important resources of the 21st century. The U.S. and China are already competing for control over technologies that, in the long term, could influence the functioning of entire nations.
The historian noted that modern power is increasingly tied to the ability to control information. According to him, the current technological race is taking place not only between the two major powers but also involves several large companies that possess vast amounts of global data and are creating increasingly sophisticated artificial intelligence models.
A YouGov poll also highlights the scale of the tech industry’s influence: about 70% of Americans believe the influence of the largest tech companies is excessive. Respondents also support stricter antitrust regulation and tighter controls on the collection of personal information.
Lessons from the Empires of the Industrial Age
Harari drew a parallel between the current development of AI and the 19th-century Industrial Revolution. During that period, the technological advantage of certain nations helped them build empires and establish control over vast territories. The gap in capabilities for producing weapons and other technologies gave the most developed countries a significant advantage.
The philosopher cited Belgium as an example. After industrialization, the country was able to establish a large colonial empire in Africa on the territory of the modern-day Democratic Republic of the Congo. In that era, industrial development and imperial power were, in fact, closely intertwined.
However, the empires of the past could not fully transfer economic resources to their centers. In Ancient Rome, land remained the most important resource: the wheat fields of Egypt or the olive groves of Iberia could not be physically transported to Italy. Therefore, a significant portion of economic and political influence remained directly in the provinces.
Similar limitations existed in the British Empire. Although industrial production was concentrated in Great Britain, natural resources and enterprises in other regions remained in their respective territories. It was impossible to transport Iraq’s oil fields or Malaysia’s rubber plantations to the British metropolis.
AI Has the Potential to Transform the Principle of Power Concentration
According to Harari, digital technologies are fundamentally different from the material resources of the past, as it is now technically possible to concentrate all the world’s information and code controlling key processes in just one or two countries.
This possibility, the historian believes, creates an entirely new level of concentration of power. A single artificial intelligence system is capable of operating simultaneously in different countries around the world, while its core infrastructure and management can be located in a single technology hub.
The Risk of Losing Technological Sovereignty
Harari also focused separately on the differences between AI and traditional weapons. Historically, a state could not remotely regain control over weapons after they had been sold or transferred to another party. He cited the Roman Empire as an example: once a steel sword had been sold to another people, the Romans could no longer remotely disable it. Similarly, in the 19th century, British authorities had no way to remotely disable the rifles that merchants supplied to Afghan tribes.
With artificial intelligence technologies, the situation may be different. If government agencies, militaries, or industrial enterprises use AI systems supplied from a single technology hub, the developer could potentially retain the ability to remotely restrict or completely shut down the operation of these systems.
Previously, the administration of U.S. President Donald Trump temporarily disabled leading AI models, such as Anthropic’s Fable 5 and Mythos 5, allegedly due to a threat to national security.
Furthermore, as the futurist emphasized, AI-based weapons are not the same as machine guns, since they are integrated into a cohesive network. At the same time, this characteristic raises new questions about state sovereignty amid the rapid development of AI.
Harari believes that in the age of artificial intelligence, the concept of sovereignty will become much more complex than it was during the era of industrial development. Depending on the system’s architecture, a mechanism for forcibly shutting down AI used outside a technological hub may remain in place within that hub.
As a result, states may face the need to choose between technological dependence on a major “empire” and opting out of the global race in artificial intelligence.
At the same time, Harari does not rule out the emergence of an alternative to the American and Chinese technological systems. According to him, creating such an alternative is possible, but there is little time left to make a decision, and those who wish to do so must act immediately.