Online hate is often dismissed as something users simply have to live with: “ignore it,” “don’t read the comments,” “nothing can really be done about it.” But research suggests otherwise — and it points to a deeper, more uncomfortable question: what if toxic content is not only a problem for social media platforms, but also part of what makes them profitable?
A field experiment by Beknazar-Yuzbashev, Jiménez-Durán, McCrosky and Stalinski tested what happened when toxic content was automatically hidden from users on Facebook, Twitter/X and YouTube for six weeks. The researchers found that reducing exposure to toxic content also reduced time spent on platforms, content consumption and advertising impressions (Beknazar-Yuzbashev et al., 2025).
That finding matters because advertising-based platforms make money from attention. Platforms are expected to protect users from harm, but their business model often depends on keeping people engaged for as long as possible. This creates a conflict of interest: toxic content may damage users’ wellbeing while also increasing the attention that platforms monetize.
In other words, toxic content is not healthy, safe or good for mental wellbeing. But it can be engaging — and that is precisely the problem.
That said, “engaging” is not the same as “valuable.” Toxic content can inflate the numbers on a dashboard while eroding the community behind them — a distinction we unpack in our piece on whether moderating hate actually reduces engagement.
The Experiment: What Happens When Toxic Content Is Hidden?
Beknazar-Yuzbashev and colleagues conducted a pre-registered browser extension field experiment across Facebook, Twitter/X and YouTube. Participants used an extension that could identify and hide toxic content while they browsed social media.
The results were clear: when users saw less toxic content, they spent less time on social media, consumed less content and saw fewer ads. The researchers also found that reducing toxic exposure decreased the toxicity of users’ own generated content. In other words, toxicity was not only something people consumed; exposure to it also appeared to influence what they produced online.
For an advertising-based platform, this creates a difficult trade-off. If hiding toxic content reduces time on platform, then reducing harm can also reduce revenue. The content that damages users may also be the content that keeps them scrolling.
Why Toxic Content Captures Attention
From a psychological perspective, this finding is not surprising. Human attention is strongly drawn to threat, conflict and social danger. Our nervous system evolved to notice rejection, humiliation, aggression and group conflict because social exclusion or attack could be dangerous.
This means that toxic content often grabs us faster than neutral or supportive content. A hateful comment, a cruel reply, a public argument or a personal attack can all activate attention, not because we enjoy them in a healthy way, but because the brain treats social threat as important information.
Beknazar-Yuzbashev and colleagues also found evidence that toxic content can trigger curiosity, increasing engagement even when engagement and user welfare are not aligned.
This matters because people may keep reading toxic comments even when those comments make them feel worse. They may feel angry, anxious, ashamed or overwhelmed — but still continue scrolling. In this cycle, emotional distress becomes engagement, and engagement becomes revenue.
Emotional Content Spreads Faster Online
The Beknazar-Yuzbashev experiment is consistent with broader research on emotionally charged content. Brady and colleagues analyzed hundreds of thousands of social media messages about polarizing political topics and found that moral-emotional language increased message diffusion. In their study, each additional moral-emotional word was associated with approximately a 20% increase in diffusion (Brady et al., 2017).
This does not mean all emotional content is toxic. Emotional language can also support solidarity, care, activism and social change. But it does show that social media environments often reward emotional intensity.
And toxic content is often emotionally intense. Outrage, humiliation, disgust, anger and public shaming can generate attention. In advertising-based systems, attention is not a neutral metric. It is monetized.
Engagement Is Not the Same as Wellbeing
Most major social media platforms are built around advertising. Their core economic goal is not necessarily to maximize user wellbeing, but to maximize attention, time spent, interaction and ad impressions.
Of course, platforms also care about safety, reputation, regulation and public trust. But their financial model still depends on keeping people online. If users spend less time scrolling, they see fewer ads. If they see fewer ads, the platform may lose revenue.
This does not mean that every platform wants people to suffer. But it does mean the system is not neutral. A business model based on attention will always struggle to fully prioritize mental health when distress itself can increase attention.
This conflict has also been explored theoretically. Beknazar-Yuzbashev and Jiménez-Durán argue that when harmful content increases engagement, advertising-driven platforms may have incentives to show more of it than would be optimal for users’ welfare (Beknazar-Yuzbashev & Jiménez-Durán, 2024).
That point is crucial: a user can engage more because they are informed, connected or entertained, but also because they are outraged, threatened, anxious or trapped in conflict. Engagement is not the same as wellbeing. Time spent is not the same as value.
Online Hate Is Repeated Psychological Exposure
At ZenFeed, we believe online hate should not be treated as a minor inconvenience or a simple moderation category. It is repeated psychological exposure.
When someone sees toxic comments again and again, especially comments about their body, identity, work, gender, sexuality, ethnicity, appearance, beliefs or personal life, the harm can become cumulative. A single cruel comment may hurt, but hundreds of cruel comments can change how a person sees themselves.
Research on cyberbullying and online victimization consistently links these experiences with negative mental health outcomes, including depression, anxiety, low self-esteem, emotional distress and suicidal ideation among adolescents and young people (Kowalski et al., 2014; Keles et al., 2020; Macrynikola et al., 2021). The U.S. Surgeon General’s 2023 advisory also highlighted cyberbullying, harassment and exposure to harmful content as important risks in young people’s digital environments (Office of the Surgeon General, 2023) — a risk we explore in more depth in our piece on why hate speech affects children and teenagers so deeply.
This is especially important for young people, creators, activists, public figures, athletes, women, LGBTQIA+ people, racialized communities and anyone repeatedly targeted online. The nervous system does not simply say: “This is just the internet.” The body can experience online attack as real social threat.
“Just Ignore It” Ignores How the Brain Works
People often tell victims of online hate to stop reading comments, block abusive accounts or grow thicker skin. But this advice places the burden on the person being harmed. It asks the user to become more resilient instead of asking why the environment is allowed to remain harmful.
From a clinical perspective, “just ignore it” is not enough. Humans are wired to notice social rejection, humiliation and exclusion because belonging is one of our basic psychological needs. This is especially true for adolescents and young adults, whose identity, self-image and emotional regulation are still developing.
If a young creator reads hateful comments about their body every day, those comments may become internalized. If an athlete is attacked after every competition, visibility may start to feel dangerous. If an activist receives threats or coordinated abuse, they may stop speaking publicly. Online hate does not disappear when the screen turns off.
Why Platform Tools Are Not Enough
Most platforms already offer safety features such as blocking, reporting, muting, hidden words, warning labels and automated moderation. These tools are useful, but they are often reactive and depend on the user taking action after harm has already occurred.
In many cases, the person has to see the abuse before they can report it. They have to read the comment, feel the impact, decide what to do and repeat the process again and again. This is emotionally exhausting.
For creators, NGOs, schools, mental health communities, youth organizations and public figures, manual moderation is often not enough. They need protection that is more proactive — tools that reduce exposure before toxic comments become part of someone’s emotional environment, as we explored in our piece on why comment sections turn toxic in the first place.
The Problem Is Bigger Than Individual Willpower
We should stop treating online hate as if it were only an individual resilience issue. The problem is not that people are too sensitive. The problem is that digital environments are often designed to maximize engagement, not psychological safety.
When platforms optimize for attention, they may unintentionally reward outrage, conflict, humiliation, emotional escalation, identity-based attacks and public shaming. Research on moral-emotional language shows that emotionally charged content can spread more widely, while research on toxic content shows that reducing toxic exposure can reduce time spent, content consumption and ad impressions (Brady et al., 2017; Beknazar-Yuzbashev et al., 2025).
Together, these findings point to a difficult reality: the content that harms people may also be the content that performs well. And if toxic content performs well, platforms may not have enough economic incentive to remove it aggressively. That is why we need independent, mental-health-informed layers of protection.
ZenFeed’s Position: Safety Should Not Depend on Platform Incentives
ZenFeed was built from a simple belief: people should not have to depend only on advertising platforms to protect their mental health.
If a platform has a financial incentive to keep users engaged, then users, creators and organizations need tools whose primary mission is not ad revenue, but protection. ZenFeed helps creators, organizations and communities reduce exposure to online hate and toxic comments.
Our goal is not to create censorship. Our goal is to reduce harm. There is a difference between disagreement and abuse, between criticism and humiliation, between debate and dehumanization, and between free expression and repeated psychological attack.
Moderation Is Mental Health Protection
Comment moderation is often discussed as a technical process, but it is also mental health protection. Every comment section has an emotional climate. It tells people whether they are safe, whether they belong, whether they can speak and whether they will be attacked.
For creators, this climate affects whether they continue showing up online. For young people, it affects how they see themselves. For organizations, activists and NGOs, it affects whether their community becomes supportive or hostile. A safer comment section is not just cleaner. It is psychologically safer.
Toxicity Should Not Be the Cost of Being Visible
One of the most damaging ideas online is that hate is simply the price of visibility. Creators, public figures, women, activists, athletes and young people are often expected to tolerate abuse as if it were an unavoidable part of being online.
But this logic normalizes harm. Visibility should not require psychological exposure to abuse. Young artists should not have to disappear to protect their mental health, teenagers should not have to absorb body-shaming as part of growing up online, and creators should not have to read hundreds of cruel comments to stay connected to their audience. Online hate should not be the cost of participation.
What This Means for Organizations
The findings from Beknazar-Yuzbashev and colleagues are especially important for organizations that work with young people, mental health, education, creators or vulnerable communities.
If toxic content keeps people engaged, then we cannot assume that platforms will solve the problem completely on their own. Organizations need to consider whether their communities are being exposed to harmful comments, whether young people are reading toxic replies before adults notice, and whether they are relying too heavily on platform-level moderation. For NGOs, schools, public health initiatives, creators, clinicians and brands, online safety is no longer optional. It is part of duty of care.
Conclusion
If we want healthier digital spaces, we need tools and systems that are not built only around attention. Success should not be measured only by time spent, comments, impressions, clicks or ad views. It should also be measured by harm reduced, toxic exposure prevented, moderation burden lowered and communities protected.
The experiment by Beknazar-Yuzbashev and colleagues shows why this issue cannot be left only to advertising platforms. If reducing toxic content means users spend less time online and see fewer ads, then platforms face a structural conflict between user wellbeing and revenue. That is why ZenFeed exists: to add a layer of protection, reduce harmful exposure, support safer communities and make digital spaces less hostile and more human.
Toxic content may be profitable. But people’s mental health should not be the price.
Sources
- CESifo Working Paper No. 11644. Beknazar-Yuzbashev, G., Jiménez-Durán, R., McCrosky, J., & Stalinski, M. “Toxic Content and User Engagement on Social Media: Evidence from a Field Experiment.”
- AEA Papers and Proceedings. Beknazar-Yuzbashev, G., & Jiménez-Durán, R. “A Model of Harmful Yet Engaging Content on Social Media.”
- PNAS. Brady, W. J., Wills, J. A., Burkart, D., Jost, J. T., & Van Bavel, J. J. “Emotion Shapes the Diffusion of Moralized Content in Social Networks.”
- Psychological Bulletin. Kowalski, R. M., Giumetti, G. W., Schroeder, A. N., & Lattanner, M. R. “Bullying in the Digital Age: A Critical Review and Meta-Analysis of Cyberbullying Research Among Youth.”
- International Journal of Adolescence and Youth. Keles, B., McCrae, N., & Grealish, A. “A Systematic Review: The Influence of Social Media on Depression, Anxiety and Psychological Distress in Adolescents.”
- Computers in Human Behavior Reports. Macrynikola, N., Auad, E., Menjivar, J., & Miranda, R. “Does Social Media Use Confer Suicide Risk? A Systematic Review of the Evidence.”
- U.S. Department of Health and Human Services. Office of the Surgeon General. “Social Media and Youth Mental Health: The U.S. Surgeon General’s Advisory.”