There is a distinction the system would rather we not make. Between knowing and learning. Between accumulating and transforming. Libraries are full of people dead on the inside, as the saying goes, and it is true, but there is something more precise to say. In the information age, that inner death has become profitable. This is not a metaphor. It is a documented business model, refined over decades, applied on a global scale. Ignorance is not the void left when education fails. It is, more often than not, the product being sold.
The Paradox of the Informed Century
Never before had humanity had access to so much knowledge in so little time. The distance between a question and an answer is now measured in milliseconds. Anyone with a phone can consult scientific studies, access historical archives, read simultaneously a Nobel laureate and a farmer documenting the destruction of his land. The potential is real. What happens to that potential is another story.
What we call the information age has produced, in parallel, an epidemic of fragmented thinking. Not by accident. The platforms that distribute this knowledge are not designed for people to understand. They are designed for people to come back. These are different things, and the confusion between the two is the source of their power. The user who learns slowly, who verifies before sharing, who changes his mind after reading something that contradicts his certainties, is, in strictly economic terms, an unprofitable user. He does not generate the kind of engagement that feeds the algorithm. He does not react at the speed that converts indignation into a click and a click into advertising revenue.
True learning, the kind that transforms, the kind that involves a small interior death of past illusions, is exactly what the model does not want.
“Doubt Is Our Product”
In 1969, an executive at Brown and Williamson, one of the major American tobacco companies, wrote in an internal memo a phrase that historian Robert Proctor would convert decades later into the axis of an entire discipline. The phrase was this. Doubt is our product. The goal was not to convince the public that tobacco was harmless. It was more efficient than that. It was enough to keep alive the feeling that science did not have the final word, that controversy existed, that the subject was complex. Cultivated ignorance does not need to deny the truth. It needs to make it appear uncertain.
Proctor called this phenomenon agnotology, the study of manufactured ignorance. His thesis, developed in Agnotology. The Making and Unmaking of Ignorance (Stanford University Press, 2008), is unambiguous, ignorance is not simply the absence of knowledge. It is, more often than not, the result of a deliberate political and cultural struggle. It has a history, a geography, financiers. When the tobacco industry paid scientists to sow doubt about the link between cigarettes and cancer, it was not making a mistake. It was executing a strategy. The same playbook was later replicated by the oil industry in the face of climate change. Today it operates, with minor technical updates, in the data centres of Silicon Valley.
An MIT study published in Science in 2018 analysed 126,000 diffusion chains on Twitter between 2006 and 2017. False news was 70% more likely to be retweeted than true news, and reached one thousand people six times faster. Platforms had access to this data. They continued optimising for engagement. (Vosoughi, Roy, Aral, Science, 2018)
The Concrete Market for Ignorance
In September 2021, Frances Haugen, former Facebook product manager, handed the United States Senate tens of thousands of pages of internal company documents. What they revealed was not a surprise to anyone inside Meta. The company’s own research confirmed that its algorithms amplified hatred, disinformation and politically extreme content because that type of content generated more reactions, and more reactions meant more screen time, and more screen time meant more advertising revenue. Haugen said it plainly before the senators. The company knew, and chose to change nothing.
The case of Myanmar is the same mechanism taken to its most brutal conclusion. Between 2012 and 2017, Facebook’s algorithms systematically amplified hate speech against the Rohingya minority, in a country where Facebook was, for the majority of the population, synonymous with the internet. The United Nations fact-finding mission that documented the 2017 genocide concluded that the platform had played a significant role in the spread of hatred that preceded the massacres. Amnesty International was more direct in its 2022 report. Meta’s algorithms proactively amplified and promoted content inciting violence against the Rohingya, while the company continued recording profits derived from that same engagement. More than 700,000 people fled to Bangladesh that year. Meta published its quarterly results without mention of the matter.
The global digital advertising market exceeds 625 billion euros. A report from the French Treasury (September 2025) documents that the negative externalities of the attention economy include the measurable deterioration of users’ cognitive capacities, costs absorbed by society while platforms count the profits.
What Haugen exposed and Myanmar confirms is neither an anomaly nor an isolated excess. It is the normal architecture of the business under conditions of low regulation. This is not about malevolent companies run by unscrupulous individuals. It is about a market logic applied to human cognition with the same efficiency that was once applied to land, labour and leisure time.
The True Learner Is a Problem
Let us return to the initial distinction, but from the angle the system does not promote. True learning is not accumulation. It is transformation. It involves changing one’s mind, which is one of the most difficult things a human being can do in public. It involves tolerating ambiguity, which is precisely what polarisation algorithms have learned to eliminate. It involves slowing down, which is the opposite of what attentional design seeks to produce.
The citizen who learns in this way is, for the system, a small but persistent problem. He does not share the news without verifying it. He does not react to the headline before reading the article. He does not need the dopamine of instant judgement. He generates little engagement. He leaves the app sooner. He is, in Silicon Valley terminology, a low-value user. The subversion is not in what he thinks. It is in how he processes.
This has an implication that is rarely named clearly. The educational system that produces efficient executors rather than thinking subjects is not an accidental failure of public policy. It is coherent with what the labour market, as currently structured, demands. UNESCO warned in 2025 that basic reading and reasoning skills have stagnated or declined in countries across all income levels. There are 739 million adults without basic literacy skills. The global learning crisis is not only a problem of access. It is also a problem of what is learned for, and who benefits from it serving that purpose.
The Recovery of the Resistant
This is where the story closes in on itself, and where any easy optimism becomes suspect. Because the system does not only produce ignorance. It also produces its antidote, packaged and sold at the price of a monthly subscription.
Mindfulness apps promise the full attention that social networks destroy. Critical thinking courses online are sold on the same platforms that monetise credulity. Books on the art of learning every day occupy the top spots in sales rankings while the recommendation algorithm learns to identify their buyers in order to offer them the next one. Slowness becomes an aesthetic. Depth becomes a niche market. The desire to exit the system is, almost without exception, absorbed by the system as a new consumption category.
It is not a conspiracy. It is more banal than that, and that is why it is harder to confront. The person who truly learns, who transforms slowly, who develops the rare form of humility that comes from knowing the immensity of what they ignore, continues to exist. But they exist at the margins of an economy that has learned to convert even resistance into a product.
Doubt, said the 1969 memo, is our product. Seventy years later, the product is more sophisticated. They no longer sell only the doubt. They also sell, at a differentiated price, the desire to overcome it…
G.S.
Sources
- Agnotology: The Making and Unmaking of Ignorance, Robert N. Proctor and Londa Schiebinger, Stanford University Press, 2008
- Disinformation is part and parcel of social media’s business model, The Conversation, 2024
- Facebook whistleblower Frances Haugen testified that the company’s algorithms are dangerous, The Conversation, 2021
- Myanmar, Facebook’s systems promoted violence against Rohingya, Amnesty International, 2022
- Social media platforms and the upside of ignorance, Centre for International Governance Innovation
- L’économie de l’attention à l’ère du numérique, Direction générale du Trésor, September 2025
- UNESCO in action, Education highlights in 2025, UNESCO, 2025
- How industry weaponizes science and sows doubt, MIT Press Reader
- The spread of true and false news online, Vosoughi, Roy, Aral, Science, 2018
