The importance of decentralizing the web for mental health

TL;DR:

• Major centralized platforms optimize for engagement, not well‑being. Their algorithms exploit our attentional and social biases.
• Effects associated with excessive use: increased depressive and anxious symptoms, stress, sleep disturbances, compulsive behaviours, addiction‑like symptoms, attention difficulties and social isolation.
• Ultra‑short content like TikTok and shorts: frequent instant gratification, reduced attentional control and a lower tolerance for boredom.
• Key mechanism: the dopamine loop. Likes, notifications and infinite scrolling rely on variable reinforcement similar to gambling mechanics, making it hard to put the app down.
• Why decentralize: less algorithmic manipulation, chronological feeds, human‑scale communities, more control over your data and moderation rules.
• Examples to explore: Mastodon and the fediverse (ActivityPub), Bluesky and the AT Protocol, Matrix for messaging, PeerTube for video.
• Takeaway: decentralization doesn’t solve everything, but it better aligns the tool with human needs and reduces harmful incentives.

Introduction

Today''s web is largely dominated by a few centralized platforms that control the flow of information online. On these centralized social networks (like Facebook, Instagram, TikTok, YouTube, etc.), content is often governed by algorithms optimized to capture users’ attention. However, more and more scientific studies show that this quest to maximize engagement can have harmful effects on users’ mental health. In response, decentralizing the web — i.e. a model where no single entity controls the platform and users can distribute themselves across interconnected independent servers — emerges as a promising avenue for a more ethical Internet that respects users’ well‑being. This article offers an accessible yet scientific overview of the impact of centralized social networks on the brain and mental health, and explains how decentralization and federated platforms (such as Mastodon or Bluesky) could offer healthier alternatives, drawing on the most recent studies and research on the topic.

Centralized social networks: what impact on mental health?

Several recent studies have established a worrying link between heavy use of centralized social networks and various psychological disorders. The symptoms and problems observed cover a broad spectrum.

• Depression: Problematic use of social networks is associated with higher rates of depressive symptoms. Systematic reviews have found that young people who are very active on social media show more signs of depression than others, all else being equal.
• Anxiety and stress: Likewise, anxiety and stress seem to be exacerbated by compulsive use of social media. Fear of missing out (FOMO)—that is, the fear of missing information or an event online—has been identified as a factor linked to this excessive use.
• Obsessive‑compulsive disorders (OCD) and attention deficit/hyperactivity disorder (ADHD) symptoms: Notably, social network abuse has also been correlated with higher levels of OCD and ADHD symptoms in the population. This suggests that the very design of these apps (constant notifications, multitasking, etc.) can aggravate impulsivity or rumination.
• Behavioral addiction: Scientists now speak of “social media addiction”. The term is justified by neurobiological observations: people addicted to social networks show brain patterns similar to other addictions (impaired inhibitory control, changes in reward regions, etc.). For example, neuroimaging studies compare the reduction of grey matter in certain areas (nucleus accumbens, amygdala, insula) in compulsive social media users to that observed in more classic addictions.
• Aggression, cyberbullying and social isolation: Intensive use of platforms can also increase vulnerability to online aggression and cyber‑harassment, which, combined with being alone in front of a screen, harms mental health. Paradoxically, antisocial behaviours can emerge from tools supposed to ‘connect’ us. Several studies also conclude that this problematic use has deleterious effects on real‑life relationships (smartphone‑related couple conflicts, feelings of loneliness, etc.).
• Disrupted sleep: Screens and constant solicitations also disturb rest. Heavy social network users show shorter sleep duration, longer and poorer‑quality sleep onset, and more nighttime awakenings triggered by notifications. Lack of sleep and night‑time phone use can in turn worsen anxiety and depression, creating a vicious cycle.

In summary, the symptom profile associated with excessive use of centralized social networks includes increased psychological distress (depression, anxiety, stress), compulsive and addictive behaviours, sleep disorders, and even measurable impacts on cognitive and brain function. It should be noted that these effects mainly affect problematic or compulsive use; moderate and conscious use of networks is much less likely to cause such harm. Nevertheless, the overall trend observed by researchers is worrying, hence the urgency of understanding the mechanisms at work.

The phenomenon of “brain rot”: when fast content reduces attention

A popular term — ‘brain rot’ (literally ‘rotting brain’) — has emerged to describe the pernicious effect that the ultra‑fast content of certain platforms (like TikTok or Instagram Reels) can have on our attention spans. While expressed figuratively, this concept echoes scientific observations on the cognition of heavy short‑video users.

Apps like TikTok offer an infinite scroll of very short, highly stimulating clips chosen by an algorithm to maximize appeal. This constant stimulation, rewarded by immediate content refresh, could lead to a form of addiction and alter our executive functions (i.e. our ability to concentrate, control and plan). Recent studies are beginning to confirm this phenomenon: for example, research based on electroencephalography (EEG) measurements has shown that young people with an addictive tendency to short mobile videos score lower in attentional control, associated with alterations in frontal brain activity. In other words, the more an individual develops a dependence on short videos, the more their self‑control and ability to stay focused are diminished.

Other work also points out that excessive use of TikTok‑style videos can lead to social withdrawal and difficulties in real‑life interactions. In purely cognitive terms, there is still limited evidence of a general decline in intelligence or memory due to TikTok. However, attention deficits and reduced tolerance for boredom among younger people are warning signs taken seriously by psychologists. Dopamine (see next section) plays a central role here: the brain, overstimulated by the instant gratification of these clips, then struggles to motivate itself for long or less exciting tasks. This is sometimes called ‘TikTok brain’: a brain trained to function only with quick rewards, at the expense of sustained attention.

In short, while the term ‘brain rot’ belongs to popular language, science partially confirms the intuition it expresses: frenzied consumption of short, viral content can impair our attention and control abilities, especially in young, developing minds. This echoes the broader finding of behavioural addiction to networks and its effects on the brain.

Dopamine: the neurological ‘hook’ exploited by Big Tech

To understand how social networks make their users addicted, we need to look at a key brain messenger: dopamine. Sometimes called the ‘happiness molecule’, dopamine is above all the neurotransmitter of reward and motivation. It is released when we satisfy a need or receive gratification — whether eating a favourite food, exercising, receiving a compliment… or seeing a notification on our phone.

Digital platforms have understood this well and have incorporated this neuropsychological lever into the design of their applications. Sean Parker, co‑founder of Facebook, publicly admitted that from the very beginnings of the social network, the goal was to capture as much of the user’s time and conscious attention as possible. To achieve this, Parker explains, ‘Facebook’s architects exploited a vulnerability in human psychology… Whenever someone likes or comments on a post, we give you a little dopamine hit.’ In other words, every ‘like’ or notification acts as a small dose of dopamine delivered to the user’s brain, reinforcing their urge to return to the app. This reward conditioning is no accident; it is directly inspired by behavioural psychology research on intermittent reinforcement.

Industrial psychologists and behavioural science experts have indeed been recruited by tech giants to make the user experience as addictive as possible. Several tech companies hire psychologists, neuroscientists and social science specialists to help them create products that maintain dopamine release in the user. Tristan Harris, a former design ethicist at Google, compares the infinite scroll of apps to a slot machine: ‘You pull a lever (swipe the screen), and you either get a tasty reward (...) or nothing. This unpredictability, akin to gambling, is precisely what keeps us coming back.’ Indeed, social networks are full of random rewards: unpredictable numbers of views or likes, notifications appearing at irregular intervals... ‘These rewards are what psychologists call variable reinforcement schedules, and that’s the key to getting users to check their screens continually,’ notes Professor Mark Griffiths, a specialist in behavioural addictions. This random reward pattern, long known to be the most addictive (according to B. F. Skinner’s experiments on rats), activates the same dopaminergic circuits in our brains as certain drugs or gambling. That’s why social platforms can trigger neural mechanisms comparable to those of cocaine, according to experts, causing a craving when the device is out of reach.

We can see why it’s so hard to ‘let go of your phone’: everything about the design of popular apps is intended to exploit our dopaminergic responses and keep us scrolling, clicking, liking, sharing. Startups like Dopamine Labs even offer their services to optimize these reward loops within other apps, using artificial intelligence and neurobehavioral analytics. In the face of this, voices — including former Silicon Valley executives — are being raised to denounce a form of ‘hacking’ of the human brain by Big Tech. New York Times columnist David Brooks summed it up: ‘Tech companies have figured out what sparks bursts of dopamine in the brain, and they pack their products with hijacking techniques that trick us and create cycles of compulsion.’

In summary, the centralized, commercial model of major networks encourages manipulative design practices aimed at making users dependent via the dopamine circuit. This frantic pursuit of maximum engagement comes at the expense of users’ mental well‑being, as we have seen. It is in this context that another path is emerging: that of a decentralized web, where platforms would be designed differently, with priorities other than ‘available brain time’.

Decentralizing the web: toward healthier, more ethical platforms

Decentralizing the web — and social networks in particular — means abandoning the single‑silo model controlled by one company in favour of an ecosystem of interconnected servers (often called instances or nodes) that different communities or organizations can operate. Each user can choose an instance (or create one) and still communicate with the entire federated network. This model, also known as ‘federation’, underpins initiatives such as Mastodon, Diaspora, PeerTube, Matrix and more recently Bluesky.

The hypothesis is that these decentralized networks, not relying on attention monetization through targeted advertising, can offer a more user‑respectful environment — potentially beneficial for mental health and the quality of online interactions.

According to a university study conducted at Oxford, administrators of decentralized social networks identify several key advantages to this approach:

• Safer and more inclusive online spaces: On federated platforms like Mastodon, there are many human‑scale communities, sometimes dedicated to marginalized or stigmatized groups, where members feel safe discussing topics they wouldn’t dare broach on the large traditional networks. Moderation is often stricter and tailored to the community, without the compromises imposed by giant platforms anxious not to offend a broad audience.
• Absence of algorithmic manipulation: A fundamental change is the abandonment of algorithmic feeds and clickbait. For example, Mastodon displays messages in chronological order, with no hidden algorithm or targeted advertising. Users see what the people they follow post, period. This avoids the traps of addictive recommendations (auto‑playing videos, polarizing content pushed to elicit emotional reactions, etc.). This return to a chronological timeline can reduce user stress and the urge to scroll endlessly, because the feed becomes manageable and finite again.
• Greater control over personal data and privacy: Decentralized platforms in principle give users more control over their data. On Mastodon, each instance has its policy, but many commit to not exploiting members’ data. Above all, a user can choose a trusted instance or even host their own server, giving them control over their data and their experience (for example, they can set their own moderation rules, export their content, etc.). This ability to leave one instance for another without losing your network (interoperability) contrasts with Facebook or TikTok’s closed model and can prevent the sense of digital lock‑in that is detrimental to well‑being.
• Prioritizing community quality over growth: Administrators of decentralized services emphasize a different philosophy: it is better to have a small but caring community than an unmanageable massive audience. This approach fosters healthier interactions, limits mass harassment and reduces the viral spread of misinformation or toxic content. In short, success is no longer measured in ‘time spent’ or number of impressions, but in the satisfaction of a community of users.

These assets suggest that decentralization can contribute to an online environment more conducive to mental health and higher-quality social interactions. Some researchers even speak of a potential for citizen empowerment through decentralized social networks: the user gains autonomy and is no longer subjected to the biases and constraints imposed by a centralized platform. By eliminating opaque algorithmic filters and the diktat of collecting likes at any cost, these federated spaces could reduce the anxiety of constant comparison and addiction to digital validation.

That said, let’s be nuanced: decentralization is no panacea and has its own challenges. The same studies highlight that these networks must address issues of moderation (each instance manages its own rules, which can lead to conflict or confusion) and sustainability (the costs and voluntary effort to maintain servers, difficulty in scaling). For example, maintaining a safe space while respecting freedom of expression is a complex balancing act for volunteer administrators. Furthermore, the user experience on federated applications can confuse novices: you have to choose an instance and understand how federation works, which is less user‑friendly than signing up for Instagram in two clicks. Finally, nothing completely prevents negative dynamics (addiction, disinformation) from appearing on decentralized networks, although the absence of amplifying algorithms limits them. In short, the technical framework offers better guarantees, but digital literacy and active moderation remain essential to realizing the well‑being benefits.

Examples of promising decentralized and federated platforms

Several projects embody this vision of a decentralized social web. Here are two notable examples:

• Mastodon: Launched in 2016, Mastodon is an open‑source social network operating through federation of instances. It is often compared to Twitter for its format (short messages called ‘toots’, follows, hashtags, etc.), but differs radically in its architecture. Anyone can create a Mastodon server or join an existing one and communicate with the entire fediverse (the set of instances linked via the ActivityPub protocol). Mastodon illustrates well the principles described above: no ads, no hidden algorithm, chronological feeds and communities with explicit rules. Research has shown that Mastodon allows groups who felt underserved or unsafe on the big networks to reclaim a more serene expressive space. The downside is that dispersal into many communities makes content discovery less immediate and can give the impression of less activity compared to Twitter’s nonstop feed. Nevertheless, Mastodon has proven its robustness and regularly attracts waves of users eager to escape the excesses of Twitter/X.
• Bluesky: This is a more recent project (beta launched in 2023) initiated by Jack Dorsey, co‑founder of Twitter. Bluesky offers a decentralized social network based on an open protocol called the AT Protocol, distinct from ActivityPub. Technically, Bluesky is designed to allow the existence of multiple servers (federated design) while offering an experience similar to Twitter. The user interface is familiar (feed of posts, replies, likes, reposts) but with a notable innovation: the user can choose their recommendation algorithm (algorithmic choice) or even create personalized feeds. This user control of the algorithm is an interesting idea for combating ‘filter bubbles’ and passive consumption of imposed content. Bluesky is still under development and relatively centralized in its initial phase (there is currently a main server bsky.social). However, in the long term, its open architecture aims to ensure that no single entity can dictate the rules or capture users’ data. Bluesky can be seen as an experiment to combine the best of both worlds: the ergonomics of a large network and the user sovereignty offered by decentralization. Its success remains to be seen, but the interest it has generated (more than a million sign‑ups in its first year) signals genuine public demand for alternatives to traditional networks.

Other initiatives deserve mention, such as Diaspora (one of the pioneers of decentralized social networks), PeerTube for decentralized video sharing, and Matrix for federated encrypted instant messaging. Even established players are showing interest in this approach: Meta (Facebook) has announced that its new Threads app could, in the future, adopt the ActivityPub protocol to interact with Mastodon and others, reflecting the growing appeal of the federated model. Each project brings its own take, but they all share the same underlying promise: to give users back control of their online lives instead of making them dependent on a giant platform.

Conclusion

The scientific evidence is clear: intensive and uncontrolled use of centralized social networks presents real risks to mental health — increases in depression and anxiety, sleep disorders, addictive symptoms, attention impairment, etc. These effects are partly the result of deliberate design choices by the platforms, which exploit our psychological vulnerabilities (dopamine, social needs, cognitive biases) to keep us online as long as possible.

However, being aware of these mechanisms enables us to act: as users, we can adopt healthier habits and turn towards more respectful alternatives.

Web decentralization offers a glimmer of hope by proposing a paradigm shift. By distributing power and prioritizing each person’s autonomy, it creates digital spaces that are potentially less toxic and more community‑oriented. The examples of Mastodon, Bluesky and others show that another model is possible — with unmanipulated feeds, human‑scale moderation and no frenzied race for attention. Of course, no system is perfect or automatically virtuous: we will still need to find a balance between freedom of expression, technological innovation and protection of mental health.

Ultimately, the importance of web decentralization lies in its ability to put people (and their well‑being) back at the center. It is a return to the original ideal of the Internet as a free and distributed network, serving its users rather than the other way around. Further studies will be needed to precisely measure the long‑term impact of these new platforms on mental health. But based on current knowledge, we can reasonably hope that more ethical, decentralized social networks will help mitigate the ‘brain rot’ linked to conventional social media and foster more enriching online interactions, benefiting our individual and collective psychological balance.

Sources: The statements in this article are based on scientific studies and publications (peer‑reviewed) as well as expert remarks, including: Weinstein (2023) on the effects of social networks on mental health, a 2024 study on the impact of short videos on attention, analyses by cyberpsychology experts on the dopamine loops induced by apps, and work from the Oxford Martin programme on the benefits and challenges of decentralized social media, among others. Each numbered reference links to the corresponding original source for readers wishing to explore these aspects further.