Social Media Was Meant to Connect Us

This article examines the impact of AI-driven feeds on an entire generation and explores what society must do before it's too late.

The Problem: AI-Driven Feeds, Mental Health, and Failing Privacy

Social media has become a natural part of everyday life for young Swedes — 94% use social media, and over 8 out of 10 do so daily. At the same time, an alarming increase in mental health issues among children and adolescents has been reported throughout the 2010s. Research points to intensified social media use as a contributing factor. A review from the Swedish Media Council found a link between high media consumption and more psychological problems among young people — with the strongest correlation for social media. Girls appear particularly vulnerable: among girls aged 9–12, the strongest links were found between frequent social media use and lower mental well-being. The Karolinska Institute reports similar findings — extensive Instagram and Snapchat use among middle school girls in Stockholm was associated with lower self-esteem and higher anxiety levels. The study ties young girls' increased vulnerability to anxiety specifically to a deteriorated self-image caused by the constant image stream.

Simultaneously, young people's digital privacy is being eroded. AI technology underpins the algorithms that control social media feeds — systems that collect vast amounts of personal data to maximize engagement. Every click, search, and pause is recorded to predict and influence our next online choice. The result is that young people's privacy is invaded by surveillance-based advertising and personality profiling that they neither consent to nor understand the scope of. Although legislation like the EU's GDPR has strengthened data protection, today's privacy safeguards are struggling to keep up with the AI era. The algorithms are complex and often secretive, the data flows cross borders, and corporate incentives favor collecting more data — not less. The outcome is a system where young users pay with their personal integrity to access "free" platforms. As one commentator recently pointed out: it took decades before we realized the privacy risks of social media — by then, the damage had already been done — and AI is now becoming even worse in an equally lawless landscape. While leading AI players have "scraped the internet for data" with no regard for privacy, legislation and oversight are falling behind. The result is that young people's right to privacy is being undermined by a constantly hungry data industry embedded in their social apps.

Empirical Observations: Swedish and International Findings

Cumulative research — from Swedish institutions to international studies — paints a picture of social media as a risk factor for young people's mental health. Swedish data: in recent youth surveys from the Swedish Media Council, a majority of teenagers report noticing negative effects themselves. Four out of ten young people say social media consumes too much of their time, and one in ten feels excluded or like a failure when comparing their lives to others' perfect feeds. The Internet Foundation reports that the problem is most visible among young women: digital appearance obsession hits them hardest, as they often feel they're wasting time online and that the constant focus on idealized appearances leads to feelings of failure. Jannike Tillå from the Internet Foundation notes that social media reinforces an already existing pressure around appearance: "Social media often focuses heavily on images, which means we are constantly being fed content that affects how we see ourselves. Growing up with comments on everything you do, notifications about everything that happens, and judgments on who you are — it makes us more vulnerable." The consequence is evident in the statistics: young women aged 12–34 are four times more likely than their male peers to feel less attractive because of social media. Young women also report more frequently than men that they feel lonely or left out when seeing others' lives online.

International studies confirm these trends. In 2024, WHO warned that the proportion of young people with "problematic" social media use — meaning an addiction-like usage pattern with inability to disconnect and negative everyday consequences — has increased significantly worldwide. A major WHO study of 11-, 13-, and 15-year-olds in Europe showed that more than 1 in 10 youths now exhibit signs of such problematic use, up from 7% to 11% between 2018 and 2022. These youths have worse mental health and sleep less than their peers with healthier habits. Several studies in the U.S. show similar correlations. Psychologist Jean Twenge has demonstrated that teenagers who spend five or more hours a day on social media are twice as likely to become depressed compared to peers who rarely use social media. Internal data from Facebook (revealed by whistleblowers) show that the company's own research found that Instagram worsens body image and increases anxiety in many teenage girls — insights the company chose to suppress for profit. In a recent Pew study (2025), 39% of American teens said social media makes them feel overwhelmed by "drama," and about a third feel pressure to post content that gets lots of "likes" or feel excluded by friends online. Over a quarter (27%) said they feel worse about their own lives after scrolling through others' posts. The same study found that teen girls are consistently more likely than boys to report negative effects — 25% of girls said social media worsened their mental health, compared to 14% of boys. There's no doubt the problem is global and gendered in similar ways: young women seem most affected by the psychological strain of algorithm-driven platforms.

The Effects of AI Algorithms: Self-Image, Stress, Perception of Reality, and Autonomy

What is it about AI technology in social media that triggers these effects? Several mechanisms become clear:

Manipulated Self-Image

AI algorithms prioritize visually appealing content — retouched images, filters, and a constant stream of others' highlights — which creates a distorted standard for self-worth. Young people compare themselves to unattainable ideals. This is evident in research: Karolinska's study linked extensive use of Instagram and Snapchat to decreased self-esteem in teenage girls, and Facebook's own analyses admitted Instagram contributes to body dissatisfaction in many youths. When the feed shows "an unattainable lifestyle" (as the Internet Foundation puts it), young people may develop a sense of inadequacy that erodes their self-image.

Increased Stress and Anxiety

The AI systems behind social media are designed to maximize the time we spend there — "they're built to be addictive," says Twenge plainly. This leads to young people being constantly bombarded with notifications and new content demanding their attention. Many feel pressured to always be available and to respond quickly, afraid of missing something (the infamous "FOMO" — fear of missing out). The result is chronic stress. The health effects are noticeable: nearly half of teens say social media worsens their sleep, and lack of sleep in turn increases the risk of anxiety and depression. Young people themselves report that constant conflicts and drama online take a psychological toll — 4 out of 10 in the Pew survey felt stressed by digital drama in their feed. AI algorithms also reinforce negative content (since scandals and conflict drive engagement), which can lead to a more negative and agitated emotional state.

Distorted Perception of Reality

Since algorithms filter and personalize content, young people receive a narrower view of the world. What they see is what the AI calculates they "like" or are provoked by — not necessarily a truthful representation of reality. This can lead to filter bubbles and polarization: a young person may be fed one-sided messages, extreme ideals, or conspiracy theories without balance, distorting their sense of what is normal and true. Even ordinary friends' posts can present a beautified version of life — everyone else seems successful and happy — making the young person misjudge their own everyday life as unusually gray. Studies have found that a significant portion of young people (27% in the U.S.) feel worse after comparing their lives to what they see online. AI also drives phenomena like deepfakes and hyper-realistic filters that blur the line between real and fake — something young users often struggle to detect. When reality in the feed becomes a glossy facade, the ability to orient — to understand what is real — is undermined.

Undermined Autonomy

Perhaps most fundamentally, AI undermines young people's personal autonomy and free will. Through sophisticated recommendation algorithms, user behavior is manipulated: platforms serve content in exactly the dose and order that keeps you glued to the screen. Young users may feel, "I've been scrolling for hours without wanting to" — a sign that their own will is being overshadowed by the design. The technology exploits psychological reward systems (likes, notifications, autoplay) to make itself indispensable, reducing young people's agency. It's telling that 58% of American teenagers (ages 15–17) say it would be hard to stop using social media. As President Joe Biden recently put it, tech companies act as if they're running "a national experiment on our children for profit." Young users are subject to influence and commercial manipulation on an unprecedented scale. Meanwhile, their private data is sacrificed — the sheer amount of personal information algorithms collect makes it nearly impossible to act autonomously online. Even if a teenager tries to protect their privacy (which many do — 40% of Swedish internet users avoid certain apps due to data insecurity), the default settings and fine print in social media rarely serve the user's autonomy. On the contrary, today's platforms are designed to collect as much data as possible and use AI to influence our choices — often without us noticing.

Inadequate Privacy Protection in the AI Era

Despite Europe's relatively strict data protection laws (GDPR and others), the protection of young people's digital rights is insufficient in the rapidly growing AI era. One fundamental problem is the scale and speed: AI-driven services can analyze and draw conclusions about a person's behavior in real time, based on massive data volumes. A single user's likes, searches, and pauses can be combined to predict private information — from political opinions to mental health — that was never explicitly shared. Current legislation struggles to address these indirect privacy intrusions. Consent buttons and cookie settings have become routine to click away rather than meaningful safeguards. For minors, the problems are especially urgent: although laws require parental consent for children under 13 to use social media, that rule is largely toothless and routinely bypassed. Platforms rarely verify age effectively, meaning even very young children engage in environments not designed — or legally allowed — for their age group.

Furthermore, major tech companies have shown little willingness to voluntarily limit their data appetite. The amount of data collected is increasing as AI needs more fuel — for example, Meta (Facebook) recently announced that it would use European users' growing volume of interactions and posts to train its AI models, despite massive criticism over privacy. For individual users, defending themselves is nearly impossible: the available settings are difficult to navigate, and default configurations always favor the data collector. The Internet Foundation notes that targeted ads and online surveillance generate widespread irritation and fear among users — many worry that large corporations are selling their personal data for commercial or political purposes. This indicates that the public already perceives privacy protections as insufficient. As AI now enables even more intrusive profiling (e.g., your voice, face, and every movement online being tracked and interpreted), the gap in protections becomes increasingly obvious. In short: we are using 20th-century legal tools to fight a 21st-century AI-driven surveillance capitalism. The result is that young people's right to privacy and self-determination is eroding before our very eyes.

Ways Forward: Policy Actions and Reforms

The problem is complex, but not hopeless. On the contrary, a progressive and solution-oriented approach can pave the way for healthier digital environments for young people. Here are several key action points on national and European levels:

Strengthen Digital Literacy and Resilience Among Youth

Media and information literacy (MIL) should be given a prominent place in school curricula. Young people need tools to understand how algorithms work, to develop critical thinking, and to handle the pressures of social media. Initiatives from the Swedish Media Council and similar actors to educate youth in digital self-defense — against filter bubbles and cyberbullying, for instance — must receive long-term funding and be integrated into education. Efforts to strengthen young people's self-esteem offline are also vital — as Karolinska researchers emphasized in their study: boosting youth psychological resilience (e.g., by promoting a healthy self-image) is crucial since social media likely won't disappear anytime soon.

Tighter Rules for Platforms — "Safe by Design"

At the policy level, social media companies must be held more accountable for young people's well-being. One possibility is to introduce age-appropriate design as the default: for example, accounts belonging to minors should have the strictest privacy settings by default, limited data collection, and no behaviorally targeted advertising. The EU has taken steps in this direction — under the new Digital Services Act (DSA), targeted ads to children are already banned, and large platforms must conduct risk analyses on how their services affect mental health in youth. These rules need strict enforcement and further development. When needed, sanctions must be impactful — multibillion-euro fines have proven to be a language Silicon Valley understands. One could also require platforms to offer non-algorithmic feeds as an option — e.g., a chronological newsfeed without AI-based sorting — so users (and especially parents of young users) can opt out of manipulative content delivery.

Transparency and Oversight of Algorithms

To regain control over AI technology, the "secret sauce" of platforms must no longer be entirely hidden. Independent researchers and regulators must be granted access to review how recommendation algorithms function, what data they are trained on, and what effects they have. The EU's upcoming AI Act moves partially in this direction by classifying certain AI systems as "high-risk" and imposing transparency and safety requirements. One possibility is to classify social media AI algorithms — given their impact on public health — as high-risk and subject to special regulatory standards. Nationally, Sweden could establish an institute for algorithm oversight that gathers expertise from Karolinska (psychological effects), the Data Protection Authority (data handling), and media researchers. This institute could analyze risks, inform the public, and provide the government with ongoing regulatory recommendations. Transparency also pressures companies to fix proven harmful features — no platform wants to be publicly seen as the one demonstrably harming children.

Protect Personal Data with Modernized Laws

Privacy protections need to be updated for the AI age. This could include everything from stricter requirements on data minimization (collecting only absolutely necessary user data) to giving individuals greater control over their digital profiles. For example, young people should have the right to easily view and delete what profiling an AI has done on them. Consent requirements must be supplemented with outright bans on certain types of data collection: particularly sensitive data (such as biometric facial data, health-related behavioral patterns, or children's usage habits) should not be usable at all by AI systems for content customization. Here, the EU could take the lead; it has already banned real-time facial recognition surveillance in public spaces in the draft AI Act — the next step could be banning the most intrusive online surveillance methods as well. Anonymization and privacy-by-design must become guiding principles in the development of new social media features. At the same time, oversight must be funded: national authorities like the Swedish Data Protection Authority and the EU Data Protection Board must be able to conduct regular audits of platforms' AI systems.

Invest in Youth Mental Health in a Digital World

Finally, we should treat the rising psychosocial problems among youth linked to social media as a public health issue — and respond accordingly. This means equipping healthcare and school health services to deal with problems related to digital stress, cyberbullying, and screen addiction. Karolinska and other research funders need continued support for longitudinal studies that can clarify causality and help us understand which measures are effective. The government could establish an interdisciplinary research program on "digital well-being" focused on children and youth, where medical experts, psychologists, sociologists, and computer scientists collaborate. Insights from this program should then be translated into practice — e.g., in the form of evidence-based guidelines on "healthy screen time" (not moral panic), and low-threshold support centers where young people or parents can seek guidance on social media use and mental health.

Toward a Sustainable Digital Future for the Next Generation

AI technology in social media has created a perfect storm of challenges for young people's mental health and digital privacy. The empirical evidence is clear: more and more young people are suffering due to constant comparison, pressure, and manipulative design, and their personal data has become a commodity on a scale society was scarcely prepared for.

But just as clear is this: it does not have to stay this way.

With smart, forward-looking reforms, we can tame the technology and once again center human well-being. History shows that societies can regulate harmful industry practices — from environmental toxins to traffic hazards — once the evidence becomes undeniable. Today, we have reached that point with AI algorithms in social media.

A child's right to grow up without being manipulated by hidden, data-driven forces or having their self-image crushed by unreachable ideals must weigh heavier than a tech company's right to maximize "engagement" at all costs.

The Swedish public debate is beginning to awaken to these issues, and internationally we are seeing the first steps toward regulation that considers both mental health and privacy. Implementing the necessary changes will require courage — but the stakes are high. We stand at a crossroads where we can choose a digital future that is sustainable, humane, and autonomy-supporting for our youth. The choice means taming AI with wisdom and values, so that social media becomes a tool for human growth instead of a health risk.

This is a choice we owe to the next generation. The sooner we act, the better the chances for today's young people to thrive both online and offline.

Read original on Medium ↗