What Can Digital Education Do About the Challenges Posed by Social Media?

This article was written in February-March 2026 for the "Les Essentiels d'Hermès" book collection, whose latest issue, Écrans contre écrans, has just been published. With their agreement, I'm reproducing my article here. Full citation: Derrac, L., 2026. "Que peut l'éducation au numérique face aux défis posés par les réseaux sociaux?", in Tsala Effa, D. (ed.), Écrans contre écrans, Paris, CNRS Éditions, coll. "Les Essentiels d'Hermès." The issue can be purchased on this page.

Any critical reflection on social media has to begin by daring to name the harm. What was seen, in the 2000s, as the realization of a digital utopia, the sharing of knowledge, horizontal collaboration, freedom of expression, has gradually turned into a largely toxic ecosystem, harmful both to individuals and to democracies themselves. Aaron Swartz, an emblematic figure of the open web, saw this coming as early as 2007, writing that everyone now had a voice, and the real question was who got heard.

More than fifteen years later, his fear has been borne out: the algorithms, interfaces, and business models of a handful of dominant platforms have redistributed the power to speak, not toward equality, but toward virality and profitability (and, at times, ideology). In a world facing overlapping crises, economic, social, geopolitical, ecological, and set against a backdrop of "post-truth," social media has become a genuine battleground in ongoing information wars.

The systemic toxicity of major social media platforms

Since the mid-2010s, the major social media platforms have repeatedly been called out for problems that go well beyond simple technical glitches or editorial missteps. These cases, numerous and thoroughly documented by journalistic and parliamentary investigations, NGO reports, academic studies, and internal audits, are not isolated errors. They reveal a systemic toxicity rooted in these platforms' economic models, algorithms, and very design. As the saying goes, it's not a bug, it's a feature. By prioritizing engagement, virality, and attention capture, social media produces harmful effects (psychological, social, political) that extend far beyond the individual and into structural, collective dimensions.

Facebook, for instance, has been at the center of countless scandals demonstrating the scale of the problems built into its economic and algorithmic models. The Cambridge Analytica affair (2018) revealed the mass exploitation of personal data for electoral manipulation. More gravely, Amnesty International showed in 2018 that Facebook's algorithms contributed to the spread of hate speech that helped fuel the ethnic cleansing of the Rohingya in Myanmar. These cases show how platforms, by prioritizing engagement over truthfulness or safety, while knowingly limiting moderation efforts, become vectors of structural violence.

Instagram, although often seen as a space for visual creativity and sharing, has been criticized for its damaging effects on young people's mental health, particularly teenage girls. An internal study conducted by Meta in 2021, made public by the Wall Street Journal, found that 32% of the teenage girls surveyed said Instagram made their feelings of depression worse. There's also the phenomenon of shadowbanning, the non-transparent suppression of content, which disproportionately affects users engaged in feminist, LGBTQIA+, or anti-racist causes. This mechanism, never officially documented by the platform, illustrates how quietly platforms can control the visibility of dissenting voices.

Even before Elon Musk's takeover, Twitter had design features that encouraged aggression and polarization: a minimalist interface, very short posts, an algorithmic timeline, and a binary reaction system built for virality (like/retweet), combined with a lack of effective moderation. Under Musk, the platform saw a marked rise in far-right rhetoric, often amplified by algorithms retuned to maximize engagement. This shift shows that "free speech" on these platforms is shaped not only by commercial logic, but by ideological logic too.

TikTok, with its endless scroll of videos, runs on a recommendation algorithm that is extraordinarily effective at capturing and holding attention. This model favors emotional, viral, and often normative content, especially around beauty, bodies, and mental health. Viral challenges, often presented as playful fun, can lead to risky or even dangerous behavior. TikTok's business model, built on targeted advertising and behavioral surveillance, makes it a tool of social control as much as entertainment.

The cases presented here aren't exceptions; they're converging manifestations of a single system. That system runs on a logic of surveillance capitalism: extreme personalization, attention capture, and the maximization of engagement, all at the expense of mental health, the diversity of voices, respect for users, and democratic stability. Understanding the systemic nature of this toxicity is a necessary condition for imagining effective responses, whether regulatory, technological, or educational. Because as long as these economic and algorithmic models go unchanged, any individual or educational intervention will remain, at best, a bandage on a deep wound.

Social media has become (a)social media: three weak signals worth watching

Recommendation algorithms, by personalizing content to the extreme, have led to the formation of informational and cultural bubbles (Pariser, 2011). Sherry Turkle (2011) already described internet users as "alone together": connected, yet isolated within our respective echo chambers. This fragmentation makes it harder to build a shared public, cultural, and informational space, one that remains essential to democratic life.

Contrary to the promise of horizontality, content production on social media is heavily concentrated in the hands of a small minority of influencers. A Pew Research Center study (2024), for example, found that 80% of posts on X (formerly Twitter) come from just 10% of users, who are mostly male, white, and politically conservative. This concentration reinforces existing inequalities in visibility and power. The gap between contributors and passive readers keeps widening.

The accumulation of content, the frantic acceleration of exchanges, information overload, and its close cousin, the fear of missing out (FOMO), have produced a form of digital fatigue. A British Standards Institution study (2025) found that 46% of young Britons aged 16 to 21 said they'd like to live in a world without the internet. This desire to disconnect isn't a rejection of technology itself; it's a demand to reclaim how it's used. This digital fatigue isn't unique to social media, either: we also talk about "decision fatigue" to describe the struggle to pick a movie on Netflix, or "Zoom fatigue" to describe the exhaustion of back-to-back video calls, something many people experienced involuntarily during the Covid pandemic.

All three of these weak signals seem to point to a deeper shift: social media isn't just a potential space for sociability, it's also a machine for producing isolation, hierarchy, and information fatigue. This is especially true of what Dominique Cardon (2019) calls "beacon" networks (réseaux phares) in his typology of social platforms. These are affinity-based networks where users can follow enormous numbers of people who are neither friends nor even acquaintances. This structure naturally favors the emergence of influencers, who accumulate what is, in effect, a form of social capital: thousands or millions of followers. It's also a structure that rewards self-promotion, since every post (consciously or not) is aimed at generating engagement: new followers, a like, a repost. Finally, it's a structure that can become an extraordinary, and distorted, echo chamber once it draws in two particular types of influencer: journalists and politicians.

It's no coincidence that users, especially younger ones, are expressing a desire to step back, not from technology or from social media altogether, but from the specific platforms that exhaust, divide, and exclude them. One clear symptom of this retreat is the huge popularity of WhatsApp groups (also owned by Meta, the parent company of Facebook and Instagram), Discord communities, and private messaging apps. These are what Cardon's typology calls "chiaroscuro" networks: fewer digital connections, and self-regulation among peers. This shift underlines the urgency of rethinking not just individual habits, but the technological and economic structures underpinning these platforms. Because as long as algorithms keep segmenting, ranking, and captivating us, social media will remain (a)social media.

Digital education: between depoliticized awareness-raising and technocritical emancipation

Digital education still tends to be confined to a preventive framework: raising awareness of risks, teaching "good practices." Drawing on the work of Simon Collin (2021), it too often serves an adaptive, depoliticized, and coercive purpose, one that reinforces individualism at the expense of collective engagement. Yet, in light of the analysis above, this approach is clearly insufficient. As Jacques Rancière (1987) put it, whoever teaches without emancipating only stultifies.

We believe a more ambitious conception of digital education is necessary, even essential given the current geopolitical and political context, one built on two principles: emancipation and technocritique. Digital education should equip people not only to understand the digital systems that structure their environment, but also to critique their complexity and, where possible, to transform them. It isn't just about diagnosing the problem; it's about building the capacity to act, both individually and collectively.

At the individual level, this means developing reflective, autonomous practices: fine-tuning personal data settings, disabling algorithmic optimization features (recommendations, personalized feeds), or turning to non-hierarchical interfaces such as chronological feeds. But above all, it means cultivating the ability to leave, work around, or boycott dominant platforms, not as a retreat, but as a political choice, in favor of existing ethical and decentralized alternatives (Mastodon, Pixelfed, Signal, independent forums, the IndieWeb, and so on), or ones yet to be imagined.

At the collective level, the challenge is to grasp the power dynamics underlying digital systems: the political and economic forces at play, and the tension between the public interest and private interests. A technocritical education should open up debate on possible regulations (banning targeted advertising, requiring algorithmic transparency, establishing a right to interoperability) and on other forms of political and economic intervention.

As Bernard Stiegler (1994) points out, digital technology is a pharmakon: both poison and remedy. Social media didn't create disinformation, violence, individualism, consumerism, or social inequality, but it has amplified all of them, through mass hyper-personalization. Educators have a crucial role to play here: teaching people to think against themselves, to resist their own conditioning, to break out of their filter bubbles. It isn't enough to train workers or consumers; what's needed are citizens of a digital world, critical and reflective, who contribute to the social web nearly as much as they consume it. In a world of overlapping crises, education must, at last, offer spaces and tools that let us slow down, think, and make room for creativity, cooperation, curiosity, and critique.

References

• Amnesty International, Myanmar: The Social Atrocity. Meta and the Right to Remedy for the Rohingya (Summary), 2022.
• Are, C., "The Shadowban Cycle: an autoethnography of pole dancing, nudity and censorship on Instagram," Feminist Media Studies, vol. 22, no. 8, pp. 2002-2019, 2021.
• British Standards Institution (BSI), Half of young people want to grow up in a world without internet, 2025.
• Cardon, D., Culture numérique. Presses de Sciences Po, 2019.
• Collin, S., "L'éducation à la citoyenneté numérique: pour quelle(s) finalité(s)?" Éducation et francophonie, vol. 49, no. 2, 2021.
• Derrac, L., "Éduquer au numérique d'accord. Mais pas n'importe lequel et pas n'importe comment, Partie 2: l'enseignement scolaire," 2024.
• Derrac, L., "Et si le problème des réseaux sociaux, c'étaient les phares?" 2024.
• Pariser, E., The Filter Bubble: What the Internet Is Hiding from You. Penguin Press, 2011.
• Pew Research Center, (2024). America's News Influencers.
• Rancière, J., Le Maître ignorant, Fayard, 1987.
• Stiegler, B., La technique et le temps. 1. La faute d'Épiméthée, Galilée, 1994.
• Tucker, J. A., et al. (2018). Social Media, Political Polarization, and Political Disinformation: A Review of the Scientific Literature. SSRN.
• Turkle, S., Alone together: Why we expect more from technology and less from each other. Basic Books/Hachette Book Group, 2011.