Freedom Index
Listen to the article
The Metric Is Not the Mission is a ten-part examination of how Big Tech moved from building and expanding the open internet to increasingly shaping it around its own metrics, incentives and assumptions. Across the series, the argument follows the evolution of the platform economy—from the optimism of the early internet to the growing tensions around power, prediction, geopolitics, accountability and the future of digital life.
The series will be published in two parts each week over five weeks, with each installment building on the one before it. At the end of the series, the complete essay will be brought together in a single PDF edition, providing the full argument in one place.
Part VI: The Illusion of Knowing Us
Part V explored what is lost when an open system becomes a finished one. Part VI returns to the question at the heart of the series: what happens when the ability to measure human behavior creates the illusion that human beings themselves have become fully understood?
There is an old temptation that accompanies every period of technological progress: the belief that a sufficiently complete collection of facts will eventually become indistinguishable from understanding. It is an idea that predates computers by centuries. Enlightenment thinkers imagined that careful observation might reveal the laws governing society just as Newton had revealed the laws governing motion. Nineteenth-century bureaucracies believed that censuses, maps and statistics would render increasingly complex populations legible to governments. Twentieth-century corporations discovered that markets could be segmented, modelled and predicted through ever more sophisticated forms of consumer research. Every era has entertained, in one form or another, the hope that enough information might finally dissolve uncertainty.
The digital age inherited that ambition and expanded it beyond anything previous generations could have imagined. Never before have private organizations possessed such detailed knowledge of everyday human behavior. The major technology platforms know what captures our attention, how long we hesitate before making a decision, which conversations draw us back repeatedly, what time we wake, when we travel, where we shop, what we read, which videos we abandon after twenty seconds and which ones persuade us to remain for twenty minutes. Individually, these fragments appear almost trivial. Collectively, they amount to one of the most ambitious efforts in history to observe human behavior at planetary scale.
The achievement is extraordinary. It is also deeply misleading. Observation and understanding are not the same thing. The distinction is easy to overlook because the predictive power of these systems has become genuinely remarkable. Recommendation algorithms often anticipate our preferences before we consciously recognize them ourselves. Navigation applications predict our journeys with astonishing accuracy. Streaming platforms learn our habits so quickly that their suggestions can feel almost uncanny. The practical success of these systems encourages a subtle but important assumption: if behavior can be predicted with sufficient precision, perhaps behavior itself has been understood.
And this may be the most consequential consequence of the platform age. The more of human behavior platforms learned to measure, the easier it became to believe that human beings themselves had become legible. Clicks became preferences. Networks became communities. Attention became interest. Prediction became understanding. Metrics became meaning. The platforms could see more of us than any institution in history and gradually became less capable of seeing what their measurements left out.
Yet this conclusion rests upon a philosophical confusion that extends far beyond technology. More than half a century ago, the economist and philosopher Friedrich Hayek argued that modern societies possess a form of knowledge that can never be fully centralized. Much of what people know is contextual, local and often impossible to articulate explicitly. A shopkeeper understands the rhythms of a neighborhood without reducing them to data points. A teacher recognizes the confidence of a struggling student before it becomes measurable through examination results. A parent notices subtle changes in a child’s mood that no questionnaire could adequately capture. This knowledge is not irrational. It is simply embedded within lived experience rather than abstract information.
Around the same time, the scientist and philosopher Michael Polanyi expressed the idea even more succinctly. “We know more than we can tell,” he wrote. Human understanding depends not only upon explicit facts but upon intuition, memory, culture, relationships and forms of judgement that resist codification. Much of what makes societies function exists precisely because it cannot be reduced to a formal rule.
Technology has always struggled with this distinction because computation requires representation. Before an algorithm can optimize anything, the world must first be translated into variables that can be measured. Human beings become profiles. Relationships become networks. Interests become categories. Attention becomes duration. Influence becomes engagement. These abstractions are indispensable because without them computation would be impossible. Yet every abstraction also excludes dimensions of reality that prove difficult to quantify.
The political scientist James C. Scott devoted an entire book to this problem. Modern states, he argued, simplify the societies they govern in order to make them administratively manageable. Forests become inventories of timber rather than ecosystems. Cities become grids. Citizens become statistics. These simplifications are not inherently malicious; they are necessary for governing large populations. Problems arise when institutions begin mistaking their simplified representations for reality itself. The map becomes more authoritative than the territory it was intended to describe.
Digital platforms confront a remarkably similar dilemma. Their models of human behavior are necessarily simplified because every computational system requires simplification. The question is not whether these models are imperfect; they are. The more consequential question is what happens when organizations become so successful within their own representations of the world that they gradually lose contact with the world those representations were created to explain.
One sees hints of this throughout the contemporary digital landscape. Platforms confidently predict what will retain our attention while appearing increasingly uncertain about what earns our trust. They identify emerging trends with astonishing speed while repeatedly struggling to distinguish civic participation from performative outrage. They optimize conversations according to measurable interactions while overlooking qualities such as reflection, empathy, restraint or wisdom because none of these can easily be incorporated into engagement metrics.
This should not surprise given that trust is not merely repeated interaction. Community is not simply network density and friendship is not the frequency of communication. Curiosity is not equivalent to clicking.
These distinctions may sound obvious when expressed in ordinary language. They become far less obvious once organizations begin making billions of decisions each day through computational systems that necessarily privilege what can be counted over what can only be experienced.
There is another irony here that deserves attention. For decades, Silicon Valley celebrated itself as uniquely capable of understanding people because it possessed unprecedented quantities of behavioral data. Traditional institutions, like governments, universities or newspapers, were often portrayed as slow, hierarchical and detached from ordinary life. Technology companies, by contrast, claimed to learn directly from users. Every click became feedback, every interaction generated insight and every product update reflected continuous adaptation to human behavior. For many years, this confidence appeared justified.
Today, however, one encounters an increasingly curious paradox. Never have companies measured human behavior so exhaustively, and never have so many users felt so profoundly misunderstood by the systems surrounding them. Social media platforms seem endlessly surprised by phenomena that emerge beyond the boundaries of their models: declining public trust, digital fatigue, growing skepticism toward artificial intelligence, the desire for smaller communities, renewed interest in newsletters, blogs, private messaging groups and slower forms of communication that escape the logic of algorithmic optimization.
These developments often appear puzzling only because they are interpreted through behavioral models that assume more engagement necessarily reflects greater satisfaction. Yet anyone familiar with modern cities knows that constant traffic does not indicate affection for the road network. Congestion may simply reveal that people have few practical alternatives. Likewise, continued use of social media reveals remarkably little about whether users believe these platforms enrich their lives. Behavior alone cannot answer that question because behavior is shaped not only by preference but also by dependency, habit, professional necessity and the absence of viable substitutes.
This may be the deepest misunderstanding at the heart of Big Tech’s current predicament. The companies continue to believe that they understand society because they observe so much of its behavior. Increasingly, society appears unconvinced. People do not feel recognized merely because they have been accurately profiled. They do not experience prediction as understanding. If anything, the extraordinary precision with which platforms anticipate our habits has heightened our awareness of everything they fail to perceive: our uncertainty, our changing aspirations, our moral dilemmas, our search for meaning, and our persistent desire to belong to communities that are valued for something more enduring than their ability to generate engagement.
History suggests that this moment arrives sooner or later for every dominant institution. Power often produces the illusion of comprehension. Success encourages organizations to believe that because they have mastered the mechanics of a system, they have also grasped its purpose. Yet societies are not machines, and human beings have an inconvenient habit of changing the questions they ask long before institutions notice that the answers they continue providing have ceased to satisfy.
Perhaps that is the quiet reckoning confronting Big Tech today. Not that it has forgotten how to build extraordinary technology. Rather, it has become so accomplished at modelling human behavior that it has begun to overlook the one thing behavior can never fully reveal: what it actually means to be human.
Konstantinos Komaitis, PhD, is a veteran of developing and analysing Internet policy to ensure an open and global Internet.
Read the full article here
Fact Checker
Verify the accuracy of this article using AI-powered analysis and real-time sources.