While artificial intelligence lowers the cost of answers, it raises the value of good questions. And good questions do not emerge from haste; they emerge from depth. Technology has already done its part: it has placed knowledge within reach of almost everyone. The challenge ahead is older and more demanding. It is not a matter of access, but of attention.
For centuries, luxury meant possessing what was scarce. First land, then material goods. Later, time. Today, the most valuable scarcity is less visible: the capacity for deep thought. We are not losing it for lack of intelligence; we are losing it because our environment no longer rewards it.
We are living through an obvious paradox. Never has it been so easy to access information, yet never has it been so difficult to sustain attention. Artificial intelligence provides answers in seconds to questions that once took hours. But while knowledge expands, concentration fragments. This is no accident; it is a business model.
As the American economist Herbert Simon predicted, an abundance of information creates a scarcity of attention. The writer Shoshana Zuboff revealed the exact mechanism: that attention is captured, processed, and sold. Platforms do not compete to inform better, but to retain users longer. They do not optimize for understanding, but for engagement.
And to retain users, the algorithm learns quickly: show each user what they already believe, what they already like, and what already irritates them. What began as personalization ends up being more like a mirror than a window. The result is not merely a culture of immediacy; it is a culture where exposure to differing ideas becomes statistically improbable.
The consequences are visible in public debate. Democratic conversation requires time, attention, and a certain intellectual discipline: listening to lengthy arguments, tolerating ambiguity, and distinguishing between the plausible and the merely appealing. Yet, these are precisely the capacities that the current environment erodes. The result is a progressive deterioration: fewer arguments and more slogans; less evidence and more identity; less deliberation and more reaction. The 2021 elections in Chile served as an unintended laboratory for this phenomenon. Boric and Kast reached the runoff as the most viral candidates. Centrist candidates—with their more complex proposals—vanished from the debate long before they disappeared from the ballot box. This was not due to a lack of ideas, but rather the absence of an environment capable of processing them.
This is where the relationship between depth and democracy becomes critical. As political scientist Yascha Mounk has argued, democracies do not collapse solely due to institutional failures, but also through the erosion of the cultural dispositions that sustain them. Central among these is the capacity to process complexity without retreating into easy answers. In this sense, populism is not merely an ideology.
It is an efficient adaptation to an environment lacking depth. Trump is the most studied case, though perhaps not the most significant one. What matters is not the content, but the format: brief, emotional, and devoid of subordinate clauses. It is a message designed—consciously or not—for an ecosystem that penalizes complexity. The problem is not Trump himself; the problem is that this format has become the standard against which everything else is measured.
The usual response to this is to appeal to individual conscience: if we understand what is happening to us, we can resist it. Yet, the evidence does not support such optimism. Being aware of a bias does not eliminate it. Knowing that algorithms are holding up a mirror to you does not make you step away from that mirror. The issue is not one of conscience, but of architecture. The environment is designed to defeat your good intentions.
What does change when we become aware operates on a different level: it changes what we demand. And this is where solutions grounded in real evidence prove to be structural rather than individual. France does not ask children to exercise self-control around their phones; instead, it removes the devices from the classroom and alters the environment. Finland does not simply teach what “fake news” is; it teaches how the algorithms that distribute it function—and it boasts the lowest rates of election-related misinformation in Europe. Some platforms have experimented with introducing friction—such as a pause or a question—before content is shared, and it works. Speed is the problem. Friction is its simplest antidote.
And then there is the most uncomfortable fact of all. The multinational tech company ByteDance operates two versions of the same algorithm. In China, children under 14 are limited to 40 minutes of Douyin (TikTok) a day, featuring educational content. In the rest of the world, the same algorithm has no limits. Tristan Harris, a former Google engineer, described it with surgical precision: it is as if they manufactured a “spinach” version for domestic use while exporting the “opium” version to the rest of the world.
The only state that has effectively limited its children’s access to the algorithm did so out of a love for control, not a love for depth.
See source
