The Attention Economy and Manufactured Anxiety
You told yourself it was just for a minute. A quick check of the news app, a brief pass through the feed, perhaps a scan of what everyone else was doing with their Tuesday morning. An hour later, your jaw is tight, your shoulders are in your ears, and you have absorbed a dense inventory of disasters, provocations, and curated evidence that other people’s lives are proceeding rather better than your own. You didn’t seek out this emotional state. You don’t particularly enjoy it. And yet — and this is the thing worth examining carefully — you will almost certainly do it again tomorrow. The fact that you will is not an accident, a character flaw, or a failure of willpower. It is, from the perspective of the companies whose platforms you were using, the system is working exactly as designed. Welcome to the attention economy: an industrial architecture built on the discovery that human anxiety, more reliably than almost any other emotional state, keeps people scrolling.
The term “attention economy” was coined by psychologist and Nobel laureate Herbert Simon in 1971 and developed into a full theoretical framework by Michael Goldhaber in the mid-1990s, but its contemporary significance was anticipated by neither of them. Simon observed that information abundance creates attention scarcity — that in a world producing more content than any person can process, the limiting resource is no longer data but human cognitive focus. What Simon could not have predicted was that the same observation would become, four decades later, the founding principle of the largest industry in human history. Global digital advertising revenue exceeded $700 billion in 2025, the overwhelming majority of it generated by platforms that offer their services free of monetary charge and finance themselves entirely through the sale of human attention. Every minute you spend on these platforms is the product being sold. Every advertisement you see is a sale being completed. The transaction is real; only the currency is unfamiliar.
There is no conspiracy here; the attention economy has learned that negative emotion is more effective than positive emotion at driving engagement. It’s an evolutionary misfire, and to understand why, we need to make a quick diversion into the architecture of the human brain. The limbic system is the part of the brain that processes threat, fear, and urgent survival information. It is much older and faster than the prefrontal cortex, where rational evaluation and deliberate decision-making are housed. The amygdala, the brain’s threat-detection centre, fires in response to alarming information before the rational mind can decide whether the alarm is warranted. UC San Diego psychiatry professor Susan Tapert, one of the lead investigators on the Adolescent Brain Cognitive Development study, describes this as a negativity bias with deep evolutionary roots: “Negative images and news tend to spark more brain activity than positive information. Historically, being alert to dangers like predators or conflict meant a better chance of survival.” The brain that learned to pay attention to threatening information outlived the brain that did not. That attention bias, which served our ancestors well on the savannah, now makes us disproportionately reactive to the specific kind of content that social media algorithms have learned to amplify.
Platform engineers didn’t design for anxiety consciously. They designed for engagement, and engagement data returned the same answer repeatedly: outrage, fear, disgust, and moral shock generate more clicks, longer sessions, more shares, and more return visits than contentment, beauty, or quiet joy. A study published in Nature Human Behaviour found that moral and emotional language in social media posts increases their spread by up to twenty percent per word. Research on X’s (formerly Twitter’s) recommendation algorithm confirmed that content evoking negative moral emotions travels significantly further and faster through networks than neutral or positive content. The algorithm did not decide that fear........
