How Meta built the scroll that never ends

You know the drill: you open Instagram “for a minute” and suddenly you’re an hour deep into Reels you didn’t even know you wanted to watch. It feels almost beyond your control – and it is. Meta’s family of apps (Facebook, Instagram, WhatsApp) are engineered around features like infinite scrolling, autoplay videos, likes, and notifications precisely because they keep us hooked. Regulators and experts now openly call these designs “addictive.” In fact, the European Commission recently warned that things like endless feeds, autoplay clips and highly personalized recommendation algorithms push users into “autopilot mode” and fuel “compulsive use”. (One EU study found teens often spend hours at night glued to Instagram, despite default limits – limits they easily ignored.) This is no accident. An EU report noted that Meta quietly builds in these features knowing they maximize screen time. Even parents’ attempts to clamp down via screen-time tools or privacy switches often fail against the firm’s “autoplay and infinite scroll” by default.

Internally, Meta’s own employees have admitted what the outside world is just realizing. Leaked slides and emails – now entering court records – show engineers and managers admitting that young users feel like they’re addicts. One internal message bluntly said, “Instagram is a drug, y’all. We are basically pushers.” Another explained that it’s no accident teens have an “addict narrative” about their usage (even as Meta forbids staff from using the word “addiction,” preferring the milder “problematic use” for legal cover). A former Instagram executive conceded under oath that the best way to drive engagement is to make the feed almost irresistible – which literally meant triggering habits like checking your phone under the desk in class. In short, the strategy is clear: use game-like tricks and data-fueled algorithms to keep you locked on the screen.

This design pays off hugely for Meta’s business. Studies show platforms profit immensely from teens: one analysis estimates that U.S. children (under 18) generate nearly $11 billion per year in ad revenue across social apps. That revenue incentive means young users are worth protecting only insofar as they stay engaged. And because Meta’s business model depends on targeted advertising, the company invests heavily in mining every click, scroll, and tap. In practice, that means everything you do on Facebook or Instagram feeds into its ad engines. Recent updates even use private AI chats as fodder for ads. As of late 2025, Meta quietly began scanning users’ conversations with its new chatbot to sell ads. In Meta’s words, if you tell its AI assistant about hiking, it learns you like hiking – and you’ll start seeing hiking boots in your recommendations and ads. In other words, your “private” chats are just another data goldmine for advertisers. Privacy-focused analysts note that this gives Meta access to our “unfiltered stream of personal thoughts” – information far more intimate than likes or follows – which can be used to detect when you’re bored, lonely or vulnerable and hit you with the perfect ad at the perfect moment. (Meta’s spin is that this will improve your experience, but the effect is that even our innermost queries to AI get logged and monetized.)

Meta insists it’s only tailoring services for our benefit. The company’s October 2025 announcement candidly stated they’ll start using your AI interactions to “personalize the content and ads you see,” promising that chatting about a hobby will surface related posts and ads. Officially, you remain “in control” via ad preferences or notification settings. But in practice, those controls are flimsy – ad options reset if you click on something new, and users report it’s nearly impossible to block all targeting. Regulators in Europe have challenged Meta on this: the EU’s privacy watchdog urged that if Facebook or Instagram offers a paid, no-ad tier, it must also give a free alternative that doesn’t use behavioral profiling. The principle is the same as placing a speed bump in the highway of engagement – something the company has so far resisted.

Meanwhile, Meta’s data collection isn’t limited to the apps you log into. Trackers on millions of websites let the company build shadow profiles even for non-users. Experts note that deleting your Meta accounts is the only surefire way to cut its reach – even then, your web browsing can leave traces. In short, Meta can keep selling you more content and ads long after you close the app, because it has dug into nearly every corner of your online life.

So what’s the harm in all this engagement? Critics argue the payoff has come at a real cost to users’ wellbeing. Public health experts have warned for years that excessive social media use is linked to anxiety, depression and sleep loss in teens. Indeed, the U.S. Surgeon General’s 2023 advisory on youth mental health bluntly notes that teenagers who spend more than three hours a day on social media face double the risk of mental-health problems like anxiety and depression – and a startling 46% of teens report these platforms worsen their body image. And that’s just average usage. Many heavy users start as children, drawn in by games and videos. One teen’s mother told The New Yorker that by high school her daughter was constantly posting to Instagram and thrilled by every like – even feeling a “buzz” of validation from them – but simultaneously became more anxious and isolated. Stories like hers are growing too common.

Related /

Inside Meta, internal studies have quietly confirmed these concerns. In one unreleased experiment (“Project Mercury”), scientists found that when users deactivated Facebook for a week, they reported lower levels of loneliness, anxiety and depression. This was so startling that staff privately warned management it was like tobacco companies hiding research on cigarettes. But the company shelved the study, calling its findings tainted by bad press. Publicly, Meta continues to claim it “listens” to experts and works on teen safety. Yet according to prosecutors, Meta’s own legal filings reveal the firm did not act on many known risks. In a landmark New Mexico case (Mar. 2026), a jury found Meta willfully harmed children’s mental health and hid evidence of dangers. The court heard that Meta made misleading claims and “unconscionable” trade-offs, preying on kids’ inexperience for profit. (In practical terms, jurors were told, Meta designed Instagram around keeping kids on the app even during class, delivering content “like you’re in a school classroom and they have a slot machine” – the more time spent, the more money earned.)

These legal crackdowns have started to bite. In California (Mar. 2026), a Los Angeles jury unanimously found Meta (and Google’s YouTube) negligent for an addicting design that harmed a 20-year-old user. The verdict awarded millions in damages, with jurors explicitly citing “infinite” feeds, autoplay videos, and push notifications as features meant to “hook” young users. And just before that, New Mexico ordered Meta to pay $375 million after finding the company violated state consumer-protection law by prioritizing growth over kids’ safety. Although Meta vows to appeal and points out that peer problems (like home life) also affect teens, the message from these cases is unmistakable: courts and regulators are treating social platforms less like neutral utilities and more like products whose hidden dangers can’t be ignored.

Through all of this, Meta portrays itself as a force for connection and says it continually updates safety tools. In reality, the company is constantly evolving its products to sustain engagement – now with sophisticated AI tools and virtual-reality experiments on the horizon. Every new feature risks adding another layer to the feedback loop: from TikTok-style Reels and Stories (short video loops designed to scroll endlessly), to AI-powered video feeds and generative chatbots, to wearable “smart glasses.” Each innovation promises user value, but behind the scenes it also plugs into Meta’s machine-learning systems. In 2026, Meta introduced Vibes, a TikTok-like feed generated by AI, and even moved to analyze voice-chat and image-prompt data to recommend content. The company says it’s all about “relevant” personalization. But as critics observe, it really means you’re feeding its algorithms more ammunition to grab your attention (and advertisers’ dollars) than ever before.

The result is an industry at a crossroads. Policymakers are debating new rules, parents are setting time limits, and young people themselves are often torn: they crave the social buzz and sense of community these apps give, yet many also experience cycles of guilt and anxiety that they can’t fully shake off. As one former Facebook executive put it under oath, Meta long resisted calling this “addiction” – preferring “problematic use” – even as all the evidence pointed that way. The EU’s new Digital Services Act and privacy laws are forcing Meta to rethink defaults (like requiring opt-in for autoplay) and even consider paid ad-free options. And after these court verdicts, many see social media as akin to an “attention casino” – powerful tools that must be regulated rather than left entirely to tech’s whims.

For now, the feed keeps flowing, and Meta’s apps remain a daily reflex for billions. But the debate over their toll is louder than ever. Every “like,” every scroll, every recommendation carries the weight of corporate design choices and real human consequences. Meta may cast itself as simply delivering what we want, but experts warn: these platforms were engineered with a very specific goal in mind – our eyes (and brains) on the screen – and untangling ourselves will take more than a simple swipe or log-out.