Both, but the harms are more severe and better-documented for heavy use—especially among adolescents—while the benefits are real but often overstated or unevenly distributed.
Documented Harms
•Mental health, particularly in teens: Multiple large-scale studies and reviews (including work by Jean Twenge, Jonathan Haidt, and others drawing on CDC, Monitoring the Future, and international data) show a sharp rise in anxiety, depression, self-harm, and suicide-related outcomes among adolescents starting around 2010–2012, coinciding with smartphone + social media penetration. Girls appear especially affected via appearance comparison, relational aggression, and sleep disruption. Experimental and quasi-experimental evidence (e.g., limiting use, natural experiments from platform outages or delayed rollout) supports a causal contribution beyond pure correlation. Heavy daily use correlates with worse outcomes; passive scrolling is worse than active use.
•Addiction and attention: Platforms are engineered for variable-ratio reinforcement (likes, infinite scroll, notifications). This fragments attention, reduces deep work/reading, and crowds out sleep, exercise, and in-person interaction. Average daily use often exceeds 2–3 hours for young users.
•Misinformation, polarization, and social comparison: Algorithmic amplification favors outrage and extremes. Echo chambers and outrage cycles worsen affective polarization. Constant upward social comparison erodes self-esteem. Cyberbullying and harassment scale efficiently.
•Other: Privacy erosion, body-image distortion (filters), FOMO, and opportunity costs (time not spent on skills, relationships, or offline competence). For some vulnerable users it enables predation or radicalization.
These effects are dose-dependent and stronger for younger brains still developing impulse control and identity.
Documented Benefits
•Connection and community: Enables maintaining distant relationships, finding niche peers (rare diseases, hobbies, LGBTQ+ youth in unsupportive areas, diaspora communities), and rapid coordination (disaster response, mutual aid).
•Information and opportunity: Instant access to news, tutorials, research, job networks, and markets. Creators, small businesses, and activists can reach audiences without traditional gatekeepers. During COVID lockdowns it substituted for some in-person contact.
•Voice and mobilization: Lowered barriers for political organizing, awareness campaigns, and marginalized voices. Empirical examples include Arab Spring coordination (mixed long-term results), #MeToo amplification, and rapid fact-checking or fundraising.
•Entertainment and low-friction utility: Humor, education short-form, professional networking (LinkedIn), and commerce.
Benefits accrue more to adults with agency, specific goals, or those who curate deliberately. They are not evenly distributed and often come with the same attention tax.
Net Assessment
For children and young teens, evidence currently tilts toward net harm when use is heavy and unsupervised—developmental costs to mental health, sleep, and social skills outweigh typical benefits. Many researchers and some jurisdictions treat this as a public-health issue warranting age restrictions, phone-free schools, or design friction (see Haidt’s The Anxious Generation and related policy debates).
For adults, it is closer to a tool whose value depends on usage patterns: deliberate, time-boxed, purpose-driven use (messaging close ties, specific learning, professional ends) can be net helpful; habitual, high-volume, algorithmically fed consumption is more often net harmful via attention, mood, and polarization. Population-level studies show average effects that are negative or null for well-being once addiction and comparison are factored in, with high variance.
Social media is not “neutral”—engagement-maximizing design creates externalities. Individual outcomes vary enormously by personality, existing support networks, content diet, and self-regulation. Absolute claims (“it’s destroying society” or “it’s pure connection”) both fail. The strongest lever most people have is reducing passive volume, curating ruthlessly, using chronological feeds or strict timers, and prioritizing offline high-bandwidth interaction.
If you're asking about a specific age group, platform, or outcome (mental health, democracy, economy, etc.), more precision yields a sharper answer.