The Austrian-British philosopher Karl Popper (1902–1994) was a committed enemy of political Utopianism. In his major work of political philosophy The Open Society and Its Enemies (1945), written in exile in New Zealand during World War II, Popper argued that any approach to politics motivated by a vision of the perfect society is likely to result in totalitarianism. Instead, he advocated the introduction of scientific method into politics—for a situation in which politicians approach societal problems in the inquiring, critical, empirical spirit with which scientists examine hypotheses. “The whole secret of scientific method,” wrote Popper, “is a readiness to learn from mistakes.”

Popper thought that the introduction of scientific methods into politics “might lead to the happy situation where politicians begin to look out for their own mistakes instead of trying to explain them away and to prove that they have always been right.” In the context of fascinating critiques of Plato, Hegel, and Marx—all of whom, he argues, helped provide the philosophical foundations for authoritarianism—Popper also rejected historicism, the doctrine that history plays out according to immutable laws, meaning that we are powerless to effect change. Against Marx, who believed that political action was useless, Popper urges us to improve the societies in which we live, not by trying to reach the “ultimate good,” but by identifying the most pressing problems in the here and now.

If we survey the political landscape 80 years after The Open Society was published, concentrating on the US and Europe, what do we see? First, that a scientific approach to politics has not become widespread. We rarely witness politicians admitting their errors or trying to learn from them: political apologies, when they’re made, tend to be for tactical reasons rather than from a desire for self-improvement. Secondly, that our modern democracies can always be improved, not to approximate Utopia more closely, but in order to maintain freedom and prevent abuses of power (the pandemic reminded us of this). And thirdly, that in the midst of the AI revolution, we could act more on the Popperian injunction to shape the kind of world we want, not look on helplessly as it renders us superfluous.

Popper’s classic work takes its title from a kind of society he argued was pioneered in Greece during the 5th century BC, by a “Great Generation” of Athenians including Thucydides, Pericles, and Protagoras (the phrase “open society” was coined by the French philosopher Henri Bergson in 1932). It was then, says Popper, that humans began to emerge from closed, tribal societies in which no distinction was made between laws of nature and man-made taboos: both were regarded as magical, unquestionable, and unchangeable. The transition into open societies that marked this distinction—in which, for the first time, societal rules were subjected to criticism and reform—brought with it what Popper called the “strain of civilization,” by which he meant “the endeavour to be rational, to forgo at least some of our emotional social needs, to look after ourselves, and to accept responsibilities.” It was, he thought, “the price we have to pay for being human.” According to Popper, this transition has now been occurring for 2,500 years. Why, then, has the scientific approach to politics yet to take hold?

The main reason—not explored in The Open Society—is surely that politicians operate in a very different environment from scientists. Scientists collaborate across borders and political divides; they form an international community guided by the pursuit of truth, united by a readiness to revise beliefs in the light of evidence. But politics is a competition, unlike science (except in the context of an arms race). Its participants are routinely locked in battles for votes, as will be the case in several EU nations next year. Admission of error is liable to be seen as a weakness and exploited by enemies, rather than welcomed as a leaping-off point by colleagues who share your objectives.

Steven Neuberg, a psychology professor at Arizona State University, has also claimed that admission of error is easier with people who know that you’re “already a moral person or that you already have good judgement”—but that politicians start from a position of distrust. Just 17% of Americans, for example, say that they trust the federal government to do the right thing “just about always” (2%) or “most of the time” (15%), one of the lowest levels seen in 70 years. In this environment, admission of error isn’t viewed as a virtue—it only confirms what most people are thinking: that you’re unfit to run the country. Politics, says the psychologist and bestselling author Steven Pinker, “is not a psychologically safe place.”

Popper insisted that the scientific approach to politics must begin with the right kind of question. He charges Plato with causing centuries of philosophical confusion by asking the wrong question, namely “Who should rule?” (Plato’s own answer was “innately superior philosopher-kings,” the basis of a political philosophy shown by Popper to be strongly totalitarian.) Popper proposed what he considered a more constructive alternative: “How can we so organize political institutions that bad or incompetent rulers can be prevented from doing too much damage?”

The effectiveness of Popper’s central question becomes clear if we look at Spain, my adopted home country. It’s been more than 1,000 days since Spain had a new state budget, due to its minority government’s lack of support in parliament. Instead, Socialist prime minister Pedro Sánchez has rolled over the 2023 budget three times, via an emergency Constitutional mechanism intended to apply for a year at most. But it was failure to pass legislation tackling Spain’s housing crisis, rather than the interminable budget deadlock, that finally prompted Sánchez to call an early election for November 29 (instead of waiting until next August). Announcing the decision on October 5, Sánchez admitted that his government had “not managed to achieve everything we set out to do”—a tactical apology buried amongst all the usual warnings about the “far right.” He didn’t elaborate on what he regards as his mistakes, or how he’s learned from them.

Sánchez’s eight-year rule, extended last time by an election he did not win, illustrates one area in which Spain’s Constitutional apparatus needs to be tweaked to prevent prolonged abuse of democracy. Fernando Navarrete, a Conservative member of the European Parliament who has called Spain’s budget freeze a “theft from citizens,” says that his party will do precisely that if it takes power after the next election.

Abuses of power also occurred during the Covid-19 pandemic, during which our open societies suffered temporary but severe regression. In politics, in the media, and amongst the general population (certainly in Europe and the UK, where some of the world’s toughest lockdowns were enforced by dubious parliamentary mechanisms), the spirit of criticism and dissent vanished or was actively suppressed. It was striking—and alarming—how quickly people acquiesced in the suspension of their most basic civil liberties, simply because governments told them that it was in their best interests. The infantile need to be told what to do was on display as never before—a remnant, perhaps, of our tribal beginnings, no doubt deeply embedded in human psychology.

It was an ominous indication of things to come when former New York Times reporter Alex Berenson’s booklet questioning the scientific basis of lockdowns—an entirely reasonable thing to do—was censored by Amazon’s self-publishing platform in June 2020. Berenson was told to remove all references to Covid-19 if he wanted the pamphlet published. “Obviously,” as he drily remarks in the introduction, “I would have a tough time

removing references to Covid-19 in a book about Covid-19” (the booklet was approved by Amazon later the same day as the attempted censorship, thanks to an online intervention by Elon

 Musk). Taking delivery of this booklet, even a couple of years after the controversy (which helped it sell more than 130,000 copies), I felt a transgressive thrill, as if I were a freedom fighter handling a piece of Cold War samizdat. The point is not whether Berenson had a good case; it is that the possibility of an open, critical discussion about lockdowns disappeared almost as soon as the pandemic began.

In the political arena, it still hasn’t happened. Yet there remain serious concerns about the legality of lockdowns, and the flimsy decrees by which European governments criminalized certain actions overnight. This is true even in Spain, where the lockdown imposed in March 2020 (during which police could fine you for walking your dog too far, or taking the long way to the supermarket) was later ruled illegal by the country’s Constitutional Court. There never has been, and obviously never will be, an apology from Sánchez for unlawfully confining an entire population to their homes. Sadly, his actions were far from sinister aberrations: according to a report published in late 2020 by Liberties and Greenpeace, more than half of EU governments failed to protect people’s right to peaceful assembly during the pandemic.

As indicated by the Berenson case, freedom of speech was also severely restricted. According to a 2021 Amnesty International report into censorship and repression during the pandemic, “the ability of everyone to engage in a debate about possible solutions and contribute to the response is a crucial element in overcoming the crisis itself… [But this] can only come about if freedom of expression and access to information are enabled.” Looking back, one can see how a Popperian approach to dealing with the epidemic would have resulted in a pleasanter intellectual atmosphere—one in which people felt free to air critical views, especially about government responses supposedly based on unquestionable science.

The reality was much nastier, as the Amnesty International report found: “Across the world, journalists, political activists, medical professionals, whistleblowers and human rights defenders who expressed critical opinions of their governments’ response to the crisis [were] censored, harassed, attacked and criminalized.” The key Popperian question to ask is: “How can we ensure that, during the next pandemic, fundamental civil liberties—including freedom of speech—aren’t abused to the extent that they were during the last one?”

Popper’s work also provides us with illuminating ways in which to look at the AI revolution. In The Open Society, he criticizes Marx for his doctrine of the inefficacy of politics in changing society (somewhat ironic, given the latter’s enormous influence). Marx believed that societal development is driven by economic laws operating independently of human agency, rendering us pawns in what he saw as the grandiose spectacle of history. Popper, however, rejected this passivity. He was a passionate believer in the power of collective and individual action to improve our shared environment. This, after all, is part of the “strain of civilization,” and it has only increased with the technological advances that have occurred since the publication of The Open Society.

But discussions about AI often cast us in a passive rather than controlling role. Laments about the increased presence of AI or robots in all areas of human life—from production lines to social media—usually present us as victims, not perpetrators: “AI is taking our jobs,” not “We are choosing robots over fellow humans.” The latter formulation prompts the key questions of why we are doing so, and whether it’s always the right decision. This is in part a sphere of moral activity, not just a practical matter of profits, competition, and productivity. In every case in which a robot is given a job formerly done by a human, that decision was made by a human, and informed by the values of their community. Such decisions contribute to the making (or unmaking) of the societies in which we live. Yet in common phrases such as “It’s just the way the world’s going” lurk traces of the passive, historicist attitude that Popper thought so dangerous.

There are signs of an AI-backlash in the customer service industry, but not necessarily due to pangs of conscience. Many companies are hiring people back into roles earmarked for automation, having found that AI agents aren’t as effective as they had hoped. According to a report by the research and advisory firm Gartner, 50% of organizations that planned a significant downsizing of their human customer service agents will abandon those plans by next year. Is this a genuine case of learning from error, of a return to valuing the human over the artificial? Or are these people being re-hired simply because AI isn’t good enough yet? If the day arrives when robots can effectively replace people in this area, the moral aspect of hiring will really make itself felt.

In an interview with the BBC earlier this year, a Chinese AI teacher summarized the brute reality of the country’s automation revolution: “If you are not needed by the world of AI and robotics,” he told the reporter in Hangzhou, “then you’ll definitely be among those eliminated.” Here, the “world of AI and robotics” is presented as an autonomous realm, not one that we are in the process of creating. If we look at the situation from a Popperian perspective, the roles are flipped back around. We can ask: “Are there areas of human life in which we should ensure that AI and robotics only ever have a minimal presence, or are eliminated?” or: “What, if any, limits should be placed on the role of AI in certain industries?” Regardless of how we answer those questions, merely asking them puts us back in control.

Popper wrote: “[I]t simply cannot be denied that we can examine thoughts, that we can criticize them, improve them, and further that we can change and improve our physical environment according to our changed, improved thoughts. And the same is true of our social environment.” And, one wants to add, of our digital environment.

Implicit in that quote is another insight—that our lives are richer (though not necessarily easier) if we do the critical thinking ourselves, rather than farm it out to the bots. Dependence on AI, however, erodes the faculties required for critical thought. A 2025 MIT study of ChatGPT as an essay-writing tool found that users “consistently underperformed at neural, linguistic, and behavioral levels,” raising “concerns about the long-term educational implications of [AI] reliance.” This highlights a fundamental truth about the AI revolution: that despite the amazing power it gives us to shape our future, it also caters to—and reflects back at us—human laziness and greed.

As early as the 1940s, Popper saw the open society changing into what he called the “abstract society,” a depersonalized world in which technology reduces the amount of physical interaction between members. Because we’re fundamentally social animals, Popper didn’t think that the totally abstract society would ever materialize—but he might have been surprised at how close we’ve come. Not entirely unpleasantly surprised, either, because he acknowledged technology’s power to open up our social horizons and, with that, the possibility of a “new individualism.” There are many things to celebrate in the open, semi-abstract societies of the 21st century; but criticism, based on the admission of error, is essential for their maintenance.