OUR ATTENTION IS THE PRODUCT.
WE ARE NOT THE BUYER.

Social-media platforms can profoundly affect the mental health and wellbeing of adolescents, with magnitudes large enough to warrant public-health-level interventions: age limits, default-off recommendations, school phone-free policies. The current product design is not a neutral tool; the data justifies treating it as a regulated environment.

“We are now experiencing one of the most consequential national experiments in real time.”

The 2010-2015 shift to smartphones + algorithmic feeds explains the simultaneous adolescent mental-health collapse across multiple Anglosphere countries. Four reforms: no smartphones before high school, no social media before 16, phone-free schools, far more unsupervised play. The mechanism is sleep, comparison, attention fragmentation, and the displacement of in-person time.

The harms come from the engagement-maximising business model, not the technology itself. Redesign incentives (subscription instead of attention auction), redesign defaults (no infinite scroll, no autoplay), redesign interfaces (calm tech, intentional friction). The industry can fix this without external regulation if it chooses to.

Treating this as 'attention' miscasts it. What is being extracted, predicted, and sold is behavioural surplus: traces of activity used to predict and modify future behaviour. The platforms are not bad for us as a quirk; they are bad for us as a logical consequence of their economic model, which is incompatible with informed democratic participation.

“We are no longer the subject. We are the raw material.”

Attention is the substrate of meaning. A fragmented attention is a fragmented life. The contemplative answer is not anti-technology but pro-discipline: meditation, sangha, monastic-style intervals away from devices. The traditions had centuries to develop tools for this; the current generation is the first that needs them at scale.

The moral-panic framing collapses several distinct issues into one (gaming, social media, smartphones, the internet). Specific products and uses have specific effects; aggregate dose-response curves are weaker than the headlines suggest. The reasonable interventions are targeted (algorithm transparency, addictive-design bans) rather than blanket (smartphone bans).

From classrooms: phones are the single largest behavioural-control issue teachers raise. From parents: device negotiation is the single largest family-conflict pattern in families with adolescents. The instruments may be debated; the lived-experience signal is large and consistent across continents.

AI is a tool that lets one human do what previously required ten. The historical pattern is that labour-saving tools eventually create more jobs than they destroy, by lowering the cost of the underlying activity (more accounting, more software, more medicine). Friction is real; net effect is broadly positive.

Twenty-first century 'automation' is unusually exposed to capture: a small number of compute-owning firms get the productivity gains while displaced workers face year-scale retraining gaps with no compensating wage growth. Without policy, this is the wrong kind of technological change.

The economic story is downstream of an existential one: as systems become more capable than humans across measurable domains, the central question is whose values they encode and whether they remain governable. The labour effects will be massive; the governance effects could be terminal.

What gets called 'AI' is a vast extractive infrastructure built on human-generated data without consent and human labour without protection. The labour displaced first is also typically the labour least visible: data labelling, content moderation, frontline service. The harms are present, gendered, and racialised; the benefits are not.

From the classroom: students who use generative AI well learn faster. Students who use it as a substitute for thinking lose the practice of thinking. The pedagogical task is not banning the tool but making the practice of un-aided cognition still happen alongside it. Many institutions are not equipped to do that.