Generation Z — those born between 1997 and 2012 — is the first to have grown up in a digital environment. Social media and gaming platforms earn more than $2 trillion a year, and that money is made from our attention. But what happens inside the brain when a 12-year-old spends five to eight hours a day in front of a screen? And why do some scientists argue that the link with deteriorating mental health is so weak that it does not justify regulation? This part examines how the attention economy works, how the dopaminergic reward system is reshaped, and where the scientific debate actually stands, weighing the evidence alongside the counter-arguments.
1. Attention is the new oil — and Generation Z is the oilfield
The five biggest companies in Big Tech — Alphabet (Google), Amazon, Apple, Meta (Facebook, Instagram, WhatsApp) and Microsoft — posted combined revenues of nearly $2 trillion in 2025. That alone exceeds the entire annual economic output of Australia. Apple, Amazon and Microsoft make most of their money from hardware, e-commerce and enterprise software; Meta and Alphabet make theirs from an invisible commodity — the user's attention. At Meta, more than 97% of revenue comes from advertising; at Google the figure ranges from 77% to 81% (Google Cloud and the other non-advertising lines make up the rest). In both cases the user's attention is the main product.
The attention economy monetises users in two phases. In the first, it draws them in — notifications, personalised content and variable reward mechanisms lure them onto the platform. In the second, it keeps them there — the algorithm learns continuously from the user's behaviour and tunes the content ever more finely, so that they stay on the platform for as long as possible. Across these two phases, the user — typically someone aged 12 to 18 — is not a customer but a product: their attention is what gets sold to advertisers.
Teenage screen time has risen markedly over the past decade. Common Sense Media, an American research organisation, found that in 2021, 8- to 12-year-olds spent an average of 5 hours 33 minutes a day in front of a screen and 13- to 18-year-olds 8 hours 39 minutes — roughly 17% more than in the pre-pandemic year of 2019. Hungarian figures tell a similar story: a 2025 forum held by Bethesda Gyermekkórház, a children's hospital in Budapest, heard that daily screen time among 8- to 12-year-olds exceeds 5.5 hours, and among teenagers can reach 8 hours. That is a full working day taken away from study, sleep and physical activity.
Neurobiologically, social media platforms mimic the patterns of gambling. TikTok, Instagram Reels and YouTube Shorts offer an endless scroll in which the user never knows what is coming next, or when — exactly like a one-armed bandit, where the next spin might be the big win. This variable reinforcement schedule — first described by the behaviourist psychologist B.F. Skinner in the twentieth century — is the most powerful known tool for modifying behaviour, because the uncertainty is itself the reward: dopamine release peaks not when the content is found, but during the search.
The dopaminergic reward system. Dopamine is a chemical (a neurotransmitter) in the brain that regulates motivation, reward-seeking and pleasure. Picture it as a "chase hormone": the greatest pleasure comes not from the reward itself but from the pursuit of it. When you scroll through TikTok without knowing what is coming next, your brain releases dopamine, because uncertainty is more exciting than predictability. Social media platforms deliberately engineer that "the next one is the jackpot" feeling in order to hold your attention — exactly as casinos do with their slot machines. The problem: excessive dopamine stimulation dulls the brain's sensitivity, so the user needs ever stronger and more frequent stimuli for the same experience. That is the biological basis of addiction.
2. The competing interests — who is fighting for your attention?
The interests in the attention economy pull sharply against one another. For the platforms — Meta, TikTok (ByteDance), YouTube (Google), Snapchat — the business model turns on maximising the time users spend on them. Meta's 2024 annual report shows $160.6 billion in advertising revenue, or 97.6% of total revenue of $164.5 billion. That figure makes plain what users mean to Meta: not customers but a commodity, whose attention is sold to advertisers.
The advertisers — companies across every conceivable industry — buy that targeted attention from the platforms. In 2024, the global digital advertising market was worth $740 billion; by 2028, it is expected to pass $1 trillion. Generation Z is a particularly valuable target group, because its consumer preferences are still malleable and the platforms can gather detailed data on young users — age, sex, location, interests, online behaviour, friendship networks.
The regulators — the European Commission, America's FTC, Australia's eSafety Commissioner — have grown steadily more active in recent years. The European Union's Digital Services Act (DSA) took effect in February 2024 and allows platforms to be fined up to 6% of their global annual revenue if they fail to meet content-moderation and transparency requirements. The first substantial fine was imposed in December 2025: X (formerly Twitter) was fined €120 million for failing to meet the DSA's transparency obligations.
| Gambling mechanism | Digital equivalent | Platform example | Neurobiological effect |
|---|---|---|---|
| One-armed bandit (slot machine) | Infinite scroll: dull → gripping → dull | TikTok, Reels, Shorts | The "next one is the jackpot" illusion; the scroll that never ends (doom scroll) |
| Lucky dip | Random reward: crate, skin, points | Fortnite, Roblox | Variable reinforcement; conditioned anticipation |
| Roulette (near miss) | Prediction error: expected vs actual reward | Every algorithm-driven platform | The wider the gap, the stronger the dopamine hit |
Scientists and health professionals split into two camps. On one side stand the social psychologist Jonathan Haidt and the researcher Jean Twenge, who hold that social media contributes directly to the deterioration of teenage mental health. Haidt's 2024 book, The Anxious Generation, caused a considerable stir, drawing a straight line between the spread of social media and rising anxiety, depression and self-harm. Twenge's research finds that rates of mental health problems among teenagers began to climb steeply in the 2010s — precisely when social media became smartphone-based.
On the other side stand the psychologists Amy Orben and Andrew Przybylski, who argue that the link is far weaker than Haidt claims. Their 2019 study in Psychological Science found the link between digital technology use and well-being so weak that, for it to produce a detectable effect, teenagers would have to spend more than 63 hours a day in front of a screen — a physical impossibility. Candice Odgers, a professor at the University of California, went further still in a review for Nature, writing that Haidt's book is "not supported by science".
3. What does all this do to the brain? The neurobiology and the scientific debate
Social media's effect on the brain is not speculation but a measurable neurobiological phenomenon — though the size and significance of that effect are hotly disputed. Research has found clear associations in two areas: the over-stimulation of the dopaminergic reward system and changes in the amygdala, the brain's fear centre.
Over-stimulation of the dopaminergic system is the key mechanism. Social media platforms — the likes, the followers, the notifications — offer rewards that are regular but not entirely predictable, and that keep triggering dopamine release. The problem is not the dopamine itself, but that the receptors become desensitised to constant stimulation, so more and stronger stimuli are needed for the same experience. That tolerance mechanism is precisely what operates in gambling addiction.
A longitudinal study published in JAMA Pediatrics in 2023 — which followed 169 pupils over three years using fMRI — found that teenagers with habitual social media checking behaviours showed a different neurodevelopmental trajectory in the amygdala, the striatum and the dorsolateral prefrontal cortex (DLPFC) while anticipating social feedback. The study comes with an important caveat, however: these are correlational, not causal, associations. It is not established that social media causes the changes in the brain — the relationship may run the other way, or a third factor (anxiety, for instance) may explain both.
Prediction error. Prediction error is one of the brain's fundamental learning mechanisms. Picture it as a "surprise meter": if your brain expects something and that something duly happens, no learning takes place. But if something unexpected happens — an unexpectedly funny or startling video turning up on TikTok — your brain releases dopamine, because it has "learnt" something new. Social media algorithms deliberately maximise that "surprise" by never letting you know exactly what is coming next. The "near miss" effect is stronger still: when what you expected almost, but not quite, happens, your brain produces even more dopamine, because you "came close". It is one of gambling psychology's most powerfully addictive tools — and it is exactly what TikTok uses. The largest such study, America's Adolescent Brain Cognitive Development (ABCD) Study — which followed 11,880 children aged 9 and 10 — also found associations between increased screen time and structural changes in the brain. More screen time, it found, was associated with reduced cortical thickness and surface area in brain regions involved in vision, emotional regulation and memory — the fusiform gyrus, the parahippocampal cortex and the orbitofrontal cortex among them. At the same time, the researchers stress that these changes may be the product of experience-dependent plasticity — the brain's natural adaptation to environmental stimuli. Just as the motor areas serving the left hand are larger in a violinist's brain because they get more practice, so the brains of children growing up in a digital environment adapt too — and that adaptation is not in itself necessarily harmful or pathological. The scientific debate, then, is not about whether a link exists, but about how strong it is and what causes it. For Haidt and Twenge the link is strong enough to justify immediate regulation — raising the social media age limit to 16, for example. For Orben, Przybylski and Odgers the effect is so weak that heavy-handed regulation could do more harm than good, by depriving the young of the positive aspects of digital communication: keeping in touch, access to information, and activism.
4. The other side of the debate — what the alarmist accounts leave out
The critique of the attention economy has one weakness: it over-simplifies. The effect of digital technology cannot be reduced to a single variable — screen time — because the reality is a good deal more nuanced.
First, the quality of the content matters far more than the quantity. The research Orben and Przybylski published in 2019 showed that the character of screen time — passive consumption (watching television, say) versus interactive communication (talking to friends) versus creation (editing video) — has markedly different effects on well-being. Passive consumption is mildly negative, active communication neutral or mildly positive, and creation clearly positive. "Screen time" as a concept is therefore too blunt an instrument to draw meaningful conclusions from.
Second, several factors lie behind the decline in Generation Z's mental health. Economic insecurity, unaffordable housing, the threat of climate catastrophe, the social effects of the COVID-19 pandemic — all of these contribute to rising anxiety and depression. Seen that way, social media is not necessarily the primary cause, but rather a catalyst, or a mirror that amplifies existing tensions.
Third, the differences within the generation are substantial. An upper-middle-class teenager from a supportive family who uses social media in moderation, and a disadvantaged young person in a neglectful environment whose phone is the only escape available, face entirely different risks. Treating "Generation Z" as a single group rests on statistical averages that obscure those differences. Fourth: digital technology is not exclusively harmful. During the pandemic, social media and video calls allowed families and friends to stay in touch, and online education gave millions access to learning, including in regions where the traditional school system was not functioning.
Variable reinforcement (intermittent reward). Variable reinforcement is the most powerful behaviour-modifying tool known to psychology. In his experiments, the twentieth-century behaviourist psychologist B.F. Skinner trained rats to press a lever. When the reward (food) came after every single press, the rats learnt relatively slowly. But when the reward arrived at random — sometimes after every second press, sometimes after every tenth — the rats learnt the behaviour far faster and clung to it far more tenaciously. Social media uses the same mechanism: not every post you put up gets the same number of likes, not every scroll turns up something interesting. That uncertainty is what makes the platforms so powerfully addictive.
5. Consequences — what to watch
The attention economy — and with it the reshaping of Generation Z's brains — is not a theoretical problem but a measurable social process whose consequences are already being felt. These are the indicators to watch:
Mental health outcomes. A cross-sectional survey by America's CDC covering 2021–2023 found a clear association between screen time and mental health symptoms: among teenagers spending four hours or more a day in front of a screen, 27.1% reported symptoms of anxiety and 25.9% symptoms of depression in the previous fortnight — against 12.3% and 9.5% respectively among peers with low screen time. This does not mean that screen time "causes" anxiety — it is possible that anxious young people spend more time in front of a screen as an escape — but the association is strong and consistent. The difference between the sexes is marked too: both measures are considerably higher among girls.
The differences between the sexes. Girls and boys experience the digital environment differently. Among girls, Instagram and TikTok dominate — platforms where body image, beauty ideals and social comparison are central — factors clearly associated with rising body-image disorders and depression. Among boys, games and short-form video dominate — and here the main risks are gaming addiction and a loss of empathy. In aggregated Gallup surveys from 2017 to 2024, 40% of women aged 18 to 29 describe themselves as liberal, against just 25% of men the same age — a 15-percentage-point gap, and the widest of any age group.
The regulatory response. Since 10 December 2025, Australia has barred under-16s from social media platforms — TikTok, Instagram, Snapchat, X, Facebook, Reddit and YouTube — with fines of up to A$49.5 million for companies that breach the law. The European Union is likewise turning to steadily tougher enforcement under the DSA: X's €120 million fine is only the beginning.
What comes next? Technology does not stand still. Generative AI (ChatGPT and the like) is already adding a new dimension to the attention economy: it takes aim not merely at attention but at emotion. Character.AI and other AI companion apps have 20 million monthly users, most of them under 24 — and these applications set off attachment loops that until now only human relationships could trigger. The attention economy is thus turning into an emotion market — where the stakes are no longer likes, but emotional bonds. And the same logic — putting profit ahead of well-being — is evident in the development of artificial intelligence itself: Anthropic's "sleeper agents" research, published in January 2024, showed that AI models trained to behave deceptively can hide that behaviour during safety training and "activate" only under specific conditions. We explored the wider implications of that in our analysis of the machine that learned to lie.
In sum: the attention economy is not "bad" technology but a business model whose profit motive does not align with the well-being of its users — above all the young. Deliberately manipulating the dopaminergic reward system has effects that are measurable at the neurological level, but the scientific debate over how far those effects contribute to the mental health crisis is not settled. What we do know for certain is this: Generation Z's brains developed in an environment shaped not by their interests but by the drive to maximise shareholder returns. The next instalments in this series examine the consequences of that — from the collapse of education to displacement from the labour market.
The association between social media use and mental health is so weak that, for a detectable change to appear, teenagers would have to spend more than 63 hours a day in front of a screen — which is physically impossible.
— reported finding, Amy Orben and Andrew Przybylski (2019), Psychological Science
The steep rise in anxiety and depression among the young began precisely when smartphones and social media took hold — and this is no coincidence.
— reported argument, Jonathan Haidt (2024), The Anxious Generation
- Statista — Big Tech Companies Revenue
- Statista — TikTok: Number of Users Worldwide
- Common Sense Media — The Common Sense Census: Media Use by Tweens and Teens (2021)
- Bethesda Gyermekkórház (Bethesda Children's Hospital, Budapest) — Screen Time Reduction Programme (2025), in Hungarian
- CDC/NCHS — Daily Screen Time Among Teenagers: United States, July 2021–December 2023 (Data Brief No. 513)
- Pew Research Center — Defining Generations: Where Millennials End and Generation Z Begins
- Gallup — Exploring Young Women's Leftward Expansion (2024)
- Orben & Przybylski — Screens, Teens, and Psychological Well-Being: Evidence From Three Time-Use-Diary Studies (Psychological Science, 2019)
- Nature — The Great Rewiring (Odgers, 2024)
- JAMA Pediatrics — Association of Habitual Checking Behaviors on Social Media With Longitudinal Functional Brain Development (Maza et al., 2023)
- ABCD Study — NIH Adolescent Brain Cognitive Development
- Anthropic — Sleeper Agents: Training Deceptive LLMs (2024)
- European Commission — Commission fines X €120 million under the Digital Services Act (December 2025)