Impact of Technology on Society: An Essay

Introduction

Technology now shapes everyday social life through systems that affect how people work, learn, communicate, and understand themselves. This essay focuses on two evidence-rich cases: automation in the economy and social media in adolescent development. These cases matter because they show why “technology” cannot be judged as a single force with one predictable social effect. Automation may raise productivity and create new work in some conditions, yet it can also displace workers when institutions and demand do not absorb the change. Social media likewise cannot be reduced to either connection or harm; its effects depend on patterns of use, individual vulnerability, and the mechanisms through which online interaction becomes supportive, stressful, or risky. Technology’s impact on society is therefore best understood as conditional rather than inherently beneficial or harmful, requiring targeted governance, clinical and educational support, and equity-sensitive research instead of broad technological optimism or moral panic.

Conditional Economic Effects

Automation is often treated as a direct route to job loss, but the stronger economic evidence shows a conditional pattern: productivity gains reduce labor per unit, yet they can also lower prices and expand demand enough to preserve or increase employment. Bessen (2019) directly challenges the “automation panic” view by showing that manufacturing industries such as textiles, steel, and automobiles experienced long periods in which employment rose alongside rapid productivity growth. The key mechanism is demand elasticity. When demand is elastic, cheaper goods stimulate enough new consumption to offset labor-saving production; when demand is inelastic, the same productivity improvement can reduce jobs (Bessen, 2019). This means the social effect of automation cannot be inferred from the technology alone. The relevant question is whether productivity growth creates expanding markets that continue to require labor, or whether it arrives after demand has already been largely satisfied.

This demand-based account also explains why the same broad process can produce both prosperity and displacement over time. In Bessen’s long-run evidence, cotton textiles, steel, and automobiles followed an inverted-U employment path: jobs grew for decades as productivity lowered prices and enlarged consumption, then later declined as markets became saturated (Bessen, 2019). The scale of productivity change was enormous—cotton textile output per hour rose more than 800-fold from 1810 to 2011—yet employment outcomes depended on whether demand kept expanding with cheaper production (Bessen, 2019). This does not deny real displacement: since 1950, many textile and steel losses are linked to productivity growth without matching demand growth, with trade and other forces explaining smaller shares in those cases (Bessen, 2019). The lesson is therefore neither optimism nor fatalism, but a demand-sensitive view of automation’s social consequences.

Contextual Mental-Health Effects

Social media is a central case for understanding technology’s social impact because it is woven into adolescent life, yet that embeddedness should not be mistaken for evidence of uniform harm. Reviews describe social media as part of everyday life for an estimated 3 billion users worldwide, with adolescents and young adults among the most active users (Schønning et al., 2020). Other adolescent-focused research similarly begins from the premise that social media is integral to young people’s lives, not an optional or marginal activity (O’Reilly, 2020; Schønning et al., 2020). That social saturation makes the issue socially important, but it also complicates interpretation. Platform use changes quickly, and some evidence bases lag behind current environments, including limited coverage of newer platforms such as TikTok (Schønning et al., 2020). The right starting point, then, is not panic about exposure itself but careful attention to how adolescents use social media.

The mental-health evidence is mixed because “social media use” and “mental health” are both broad categories that conceal different mechanisms and outcomes. Schønning et al. (2020) found that most studies reported some correlation between more frequent social media use and poorer well-being or mental health, but they also emphasized that the field remains contested and difficult to interpret. This caution is important because much of the evidence cannot establish cause and effect: in their scoping review of 79 studies, cross-sectional designs were the most common, while only 16 studies were longitudinal (Schønning et al., 2020). Fassi et al. (2025) similarly note that existing studies often show small associations and conflicting results rather than a clear causal relationship. Qualitative work reinforces this complexity, with adolescents describing social media as having “good,” “bad,” and “ugly” sides (O’Reilly, 2020). The evidence therefore supports a contextual interpretation, not a simple exposure-harm model.

The strongest distinction is between crude measures of time online and more problematic patterns of engagement. Many studies have relied on frequency and duration, with one review finding that this was the most common measure of social media use, used in 56% of included studies (Schønning et al., 2020). Yet time spent often misses what adolescents are actually doing online. Boer et al. (2020), using a three-wave longitudinal study of 2,109 adolescents, found that social media use problems predicted poorer mental health one year later, while use intensity was not associated with mental health over time. Fassi et al. (2025) complicate this further: adolescents with mental health conditions did report more time on social media, but the authors treated time as a crude self-reported measure and examined qualitative engagement patterns as well. These findings shift the question from “how much screen time?” to “what kind of engagement, for whom, and under what conditions?”

Vulnerability and Risk Mechanisms

The clearest evidence of social media harm points less to all adolescents than to adolescents whose clinical profiles may make online interaction more stressful. This builds on the earlier distinction between time online and qualitatively different engagement: time is a weak proxy, but vulnerable groups may still experience more intense forms of use. In a clinical comparison of adolescents with and without diagnosed mental health conditions, Fassi et al. (2025) found that adolescents with internalizing conditions showed higher social media time, more online social comparison, and greater mood impact from online feedback than adolescents without a condition. The difference in online social comparison was meaningful enough to matter, with an effect of g = 0.54 for internalizing conditions versus no condition (Fassi et al., 2025). This pattern did not apply evenly across all diagnoses: adolescents with externalizing conditions mainly differed in time spent online, and those findings require caution because of smaller subgroup size and comorbidity issues (Fassi et al., 2025). The risk, then, is differentiated rather than universal.

The mechanisms most associated with harm—comparison, feedback sensitivity, and cybervictimization—also show why moral-panic accounts are too blunt. Reviews identify cyberbullying as a major concern, with evidence linking cybervictimization to internalizing and externalizing problems and suicidal behavior risk among adolescents (Schønning et al., 2020). At the same time, longitudinal evidence complicates a simple pathway from online exposure to later mental-health decline: Boer et al. (2020) found that social media use problems predicted later upward comparison and cybervictimization, but those processes did not mediate the observed decline in mental health. This does not make the risks imaginary; it means the causal chain is not yet settled. A further problem is interpretive bias. Schønning et al. (2020) found that about three-quarters of studies focused on pathology, while practitioners in O’Reilly (2020) often repeated negative rhetoric even while acknowledging possible benefits. Society should therefore take specific risks seriously without turning them into a generalized story of technological damage.

Targeted Social Responses

Effective responses should use technology where it can reach people, rather than treating digital tools only as sources of harm. In adolescent mental health, social media is not simply an exposure to reduce; it is also a space where young people already seek information and where support can be delivered. O’Reilly et al. (2018) found that adolescents commonly use social media and the internet to look for mental-health information, and the study concluded that, despite risks, social media can help educate and reach adolescents to promote well-being. This does not erase concern, since practitioners and adolescents describe social media as having beneficial and harmful sides, but it does challenge one-sided public narratives of pathology (O’Reilly, 2020; Schønning et al., 2020). The social response should therefore be practical rather than reactionary: use digital platforms for outreach while designing them around known risks.

Clinical and educational systems also need targeted assessment rather than broad assumptions about screen time. Earlier sections showed that the strongest concerns involve heterogeneous, often non-causal associations and that adolescents with internalizing conditions may be especially sensitive to online comparison and feedback. That pattern supports risk assessment that asks how social media is being used, not just whether it is being used. O’Reilly (2020) argues that practitioners should consider social media as a dimension of adolescent risk assessment without presuming the relationship is negative. This fits the broader evidence that adolescent mental-health burdens create challenges for education and healthcare systems, while providers are looking for effective ways to promote positive mental health (O’Reilly et al., 2018). The appropriate intervention is therefore not blanket restriction, but careful screening, digital literacy, and referral pathways for adolescents whose online experiences intensify distress.

Targeted governance must also confront the limits of the evidence used to make social decisions. Better measurement may require real-time tracking of actual use, because self-reported estimates of duration and frequency are often unreliable, but Schønning et al. (2020) note that such tracking raises obvious legal and ethical challenges. That tension matters because more objective data can improve research while also expanding surveillance risks if privacy is not protected. Equity is another limitation: gender was included in most studies but rarely analyzed as an explanatory variable, and socioeconomic and geographic differences are also underexamined (Schønning et al., 2020). Even strong recent clinical evidence is limited by context, since nationally representative English data may not generalize across regions shaped by different social and cultural conditions (Fassi et al., 2025). Technology policy should therefore pair intervention with privacy safeguards and equity-sensitive research.

Conclusion

The evidence across automation and social media points to the same conclusion: technology’s social impact depends on the conditions around its use, not on the tool alone. Automation can expand employment when productivity gains create new demand, but it can also displace workers when demand no longer grows enough to absorb change. Social media can support information-seeking, connection, and outreach, yet it can also intensify distress for adolescents facing comparison, feedback sensitivity, cybervictimization, or existing mental-health vulnerabilities. These patterns make broad technological optimism and broad moral panic equally inadequate. A more responsible social response is targeted: economic policy should focus on demand, transition support, and worker protection, while schools and clinicians should assess how young people use platforms rather than treating screen time as the whole problem. Future governance should also protect privacy and address gaps in evidence across gender, socioeconomic status, and place. Technology should be managed as a conditional social force, guided by evidence rather than assumption.