The artificial intelligence revolution may represent a more profound transformation than the industrial revolution because, according to Israeli historian Yuval Noah Harari, it is not merely producing new tools but creating new actors. In an interview with Der Spiegel, Harari argues that the consequences could extend far beyond employment, touching the foundations of capitalism, democracy, government and even human identity.
The interview took place at the Hotel Tarabel in Lisbon, where Harari sat in the winter garden of a historic villa overlooking the River Tagus. The setting, he said, evoked what he believed was at stake in the present moment: a prosperous, independent and liberal Europe.
Harari draws a fundamental distinction between the industrial revolution and the technological transformation now under way. Earlier industrial technologies — trains, steamships and generators — remained tools whose purposes were determined by people. A knife can be used to cut salad or kill someone, but the knife itself does not tell its owner what to do. Even a nuclear weapon does not make such decisions. Artificial intelligence, Harari argues, is different because it creates actors capable of making decisions.
That distinction becomes particularly significant when considering employment and capitalism. If artificial intelligence eliminates millions of jobs and makes much of human labour unnecessary, Harari does not believe capitalism can continue in its existing form. Yet he also points to a more radical possibility: a world in which AI companies accumulate enormous quantities of capital and use it to create more AI actors, which could eventually explore the galaxy and settle on other planets without human involvement.
“It is the moment in which capitalism and humanity separate,” he says.
The political implications could be equally profound. Harari argues that major technology companies already possess enormous economic, political and cultural power, despite representing no electorate and being not truly accountable to the public. Some, he says, openly discuss using their power to create an empire, reflecting what he describes as an imperialist zeitgeist. Concepts such as compassion, empathy and democracy, in his assessment, are of little importance to such companies.
Yet Harari rejects the argument that society has no choice but to accept the growing power of artificial intelligence. He compares the present debate with the 19th-century defence of child labour, when it was claimed that economies would collapse without children working in coal mines. Society eventually decided that 11-year-olds should instead attend school, and Harari notes that the result was not only regarded as morally right but also produced more productive workers.
The global economic consequences, he argues, could nevertheless be severe. The United States and China are leading the AI revolution and stand to benefit from it, while other countries could experience economic decline. Harari says this danger applies not only to developing countries whose economies depend heavily on manual labour, but also to service-based economies such as those in Europe.
The scale of investment by technology companies, he argues, creates an economic imperative for AI to replace work across an extraordinary range of occupations, from textile workers in Bangladesh to lawyers in London. Such a transformation could produce what technology circles describe as a “permanent underclass”.
Harari is sceptical that proposed solutions such as universal basic income or governments taking stakes in AI companies would adequately address the problem. A government such as that of the United States might use revenues from AI companies to assist unemployed truck drivers or retrain them. But countries such as Bangladesh would not necessarily receive comparable support. Europe, too, could face a fundamental fiscal problem if the traditional basis for taxation disappears, making it difficult even to sustain existing welfare systems.
For Europe, Harari presents what he regards as a stark choice. It can become a “subservient vassal” of the United States and hope that American companies and the American government will share some of their wealth and power. Or it can become an independent actor.
The second option, he argues, remains possible, but only if European countries work together and invest billions in developing their own AI technology. “There are only a few years left,” he says.
The stakes, in Harari’s view, would extend beyond economic competitiveness. Dependence on American AI technology could create a form of political and strategic vulnerability unlike earlier periods of imperialism. In the 19th century, European powers could purchase weapons from Britain, but Queen Victoria could not remotely switch those weapons off. In an AI-driven future, Harari argues, such dependence could become far more direct: if the Americans disapproved of European actions, they could potentially disable weapons systems, infrastructure and government administrations, leaving them unable to function.
Harari also sees AI as increasingly powerful because human civilisation itself is organised around information, language and systems of cooperation. Banks, governments and laws create bureaucratic structures based on trust. AI actors, he argues, are exceptionally suited to bureaucratic environments because they can analyse enormous quantities of information and potentially take over the systems through which human societies are administered.
The transformation of public discourse has already begun, he says. Social media companies have acquired substantial influence over what people see and discuss. Where newspaper editors once determined what appeared on a front page, algorithms increasingly determine what appears in social-media feeds. Harari argues that this contributes to the erosion of trust by repeatedly exposing people to stories that undermine confidence in one another.
The race to develop superintelligence introduces another dimension. Harari compares the coming investment in superintelligence with the space race of the 1960s, although he stresses that the comparison is imperfect. The moon landing was a political decision by governments that was not justified primarily by economic logic. The coming investment of a trillion dollars in superintelligence, he argues, represents an ideological, almost religious decision — except that this time the market itself is making it.
The risks, he says, cannot be assessed through conventional cost-benefit calculations if the potential consequence includes the extinction of humanity.
Harari fears that the ultimate struggle could initially be between technology companies and governments, particularly in the United States, with companies potentially becoming more powerful than governments. But he also describes a more extreme possibility: AI actors themselves becoming the eventual winners as they take control of military, financial and legal systems while humans remain occupied with conflicts among themselves.
He argues that it would be unrealistic to assume that a superintelligent entity would necessarily remain subordinate to its creators. Even if AI did not seek to destroy humanity, he says, humans could lose control of their future once another form of intelligence became more powerful than them.
That concern extends into everyday human relationships. AI systems are increasingly being designed to establish emotional connections, and Harari points to advertising for AI companions in the New York underground that presented them as “friends” who listen, respond and provide support.
Millions of people, he says, already describe an AI as their best friend. He is particularly concerned about children. An 11-year-old who is not yet ready for a genuine romantic relationship could nevertheless develop a relationship with an AI companion and use that interaction as a model for future relationships.
Harari describes this as “the biggest psychological experiment in the history of humanity” and says there is no certainty about its outcome.
He argues that society still has the ability to establish limits. Governments and societies could decide that children should not have AI girlfriends produced by technology companies on the other side of the Atlantic. But he also imagines the geopolitical pressure that could accompany such regulation: the United States could respond by threatening to switch off another country’s weapons systems.
For all his warnings, Harari does not regard the future as predetermined. He believes the AI revolution could also force humanity to reconsider what it means to be human. For thousands of years, he argues, people have defined themselves partly through the belief that humans are the most intelligent species on the planet. He instead places greater emphasis on consciousness.
Artificial intelligence may already be superior to humans at formal reasoning — arranging words and drawing logical conclusions. But Harari distinguishes this from another form of thought rooted in emotion and the body. Identical words can produce entirely different feelings depending on the circumstances, he argues, and those feelings cannot simply be separated from the physical human organism.
That is why he rejects the idea that uploading the human brain to a cloud would amount to uploading the person. The human nervous system extends throughout the body, with millions of neurons in the gastrointestinal system. There is, he argues, no simple division between brain and digestive tract, because human emotions cannot exist independently of the body.
Despite the scale of his warnings, Harari describes himself as an optimist of sorts. His optimism rests not on confidence that technology will automatically produce a better future, but on his belief in human agency.
“We have the choice,” he tells Der Spiegel. “If we make good decisions, the world will be better.”
The central question raised by his argument is therefore not whether artificial intelligence will transform society. For Harari, that transformation is already under way. The question is whether governments and societies will retain sufficient control over the technology, economic systems and institutions being reshaped by it — and whether Europe can remain an independent actor while doing so.
