In June 2026, the Complaints Commission of the Serbian Press Council considered the reporting of several Serbian news portals on a murder suspect. One headline read, “From Respected Police Officer to Executioner”.

The Commission found that some media outlets had violated the presumption of innocence. Commission member Milena Vasić warned of a consequence that is difficult to reverse: when enough media outlets label someone a murderer, a later acquittal cannot simply erase what has already remained in the public mind.

The public verdict often arrives before the court’s.

The name has been published. The photograph circulates online. Alleged details from the investigation begin to appear. Comments demand life imprisonment, lynching, dismissal, expulsion or something worse. Someone finds the family’s profile. Someone else publishes an address.

Within hours, an entire story exists about what happened, who is guilty and what kind of person the alleged perpetrator must be.

The evidence has yet to be collected.

The Constitution of Serbia is fairly clear on this point. Article 34 provides that everyone shall be presumed innocent until their guilt has been established by a final court decision. The Code of Journalists of Serbia requires the same from the media and specifically warns about the possibility of false accusations and the need to identify unproven allegations as such.

The legal standard is simpler than human behaviour.

Because when something terrible happens, a large part of the public does not first ask what can we prove?

It asks a much more immediate question:

Who is going to pay for this?

Anger reaches a verdict much faster

The desire to punish has a deep psychological logic.

Psychologist John Darley, in reviews of research on punishment, described a moral violation as an event that can provoke an intuitive and emotionally charged reaction of moral outrage in observers. Particularly when harm is perceived as intentional, the desire for retaliation against the person identified as responsible increases.

A large study published in 2024 in Scientific Reports, using data from 56 societies, also found a strong relationship between emotions and support for social punishment. People who felt more anger and disgust in response to violations of social norms were more likely to support confrontation and other sanctions, with anger being particularly strongly associated with a desire to confront the offender directly.

That does not mean anger is irrational.

If someone has been killed, beaten, raped, robbed or seriously abused, anger is an expected human response. A society that showed no reaction to violence would be equally troubling.

The difficulty comes in the next step.

The emotion that tells us an injustice has occurred can very easily turn into the belief that we already know who is guilty, why they did it and what punishment they deserve.

Between those two things lies an entire criminal proceeding.

Uncertainty is unbearable

The first few hours after a serious crime are often the worst possible moment for reliable conclusions.

Information is incomplete. Witnesses can be wrong. Videos lack context. Police possess some evidence that has not yet been made public. Journalists have fragments. Rumours, wrong names and photographs of people who may have nothing to do with the event circulate on social media.

It is precisely at that moment that there is an enormous need to give the event a clear structure.

A study published in 2023 in Personality and Individual Differences examined the relationship between the psychological need for cognitive closure and attitudes towards punishment. In a sample of 505 people in Italy, a stronger need for a quick and stable answer in situations of uncertainty was associated with more pronounced punitive attitudes.

A simple answer reduces chaos.

There is a crime.

There is a criminal.

There is a punishment.

The problem is that real cases often look far more complicated once the files are opened.

There may be several participants. The first suspect may not be the perpetrator. The offence may not legally be what the headline says it is. A video may have been edited. The initial testimony may be inaccurate. Behind individual responsibility there may be an entire chain of institutional failures.

All of that takes time.

Anger does not operate on the same deadline.

Social media makes anger look bigger than it is

Online, another problem appears: the perception of the majority.

A study published in Nature Human Behaviour in 2023 found that social media users systematically overestimate how much moral outrage other people actually feel.

In experiments and research on behaviour on Twitter, observers concluded from posts that their authors were angrier than the authors themselves reported being. The effect was stronger among people who relied more heavily on social media for political information. Such overestimation also increased the impression that social groups were more hostile and extreme than they actually were.

That changes the atmosphere in which an individual forms their own judgement.

If we open the comments and see hundreds of demands for the harshest punishment, it is easy to conclude that there is complete social consensus.

If doubt appears, it can suddenly seem almost indecent.

“How can you defend him now?”

That is one of the most effective ways to shut down a discussion.

Asking for evidence begins to be interpreted as defending the suspect. Insisting on due process appears to show a lack of empathy for the victim. Even the simple statement that we do not yet know what happened can sound like taking sides.

Moral position is then measured by the amount of anger a person is willing to display publicly.

Research from 2025 additionally found that people may judge an observer who responds with anger to a moral violation as a more moral and trustworthy person. In other words, the public display of outrage can also carry a social reward.

In such an environment, caution is far less attractive than condemnation.

The media have a culprit before the court has a case

The media also determine which facts the public will use to construct its picture of events.

Experimental research published in 2022 found that the way a report about a criminal offence is written can change readers’ assessments of guilt and appropriate punishment.

It is particularly problematic when information carrying a strong moral or political charge is inserted alongside the offence even though it is not decisive for determining criminal responsibility.

Social media accelerate this logic. A study involving more than 900 American students found an association between consuming social media content focused specifically on punishment and stronger punitive attitudes.

The authors did not claim that every such piece of content directly produces a desire to punish, but the finding points to a relationship between the media environment and the way people imagine justice.

A more recent experimental study, published in June 2026, found that even the way a short video about crime is filmed can influence punishment preferences. The study was small, with 60 participants, so its results should not be turned into a universal rule.

Still, it shows how the presentation of an event can become part of our moral judgement of it.

Tabloid media understood this long ago without reading academic papers.

“Suspect” is a weak word for a headline.

“Monster” works better.

A court case can last for years. A headline needs the click now.

Punishment is also a political product

Punitiveness is not produced only in comment sections.

A politician who promises “the strongest possible response from the state” after a serious crime is in a much easier position than a politician who explains prosecutorial jurisdiction, standards of proof and the limits of their own office.

The first demonstrates strength.

The second talks about procedure.

Research on political rhetoric and public opinion in 26 European countries, published in the British Journal of Criminology, found an association between political “law and order” rhetoric and citizens’ preference for harsher punishment.

The authors conclude that the political frame through which crime is discussed can influence public punitive preferences.

That makes the political exploitation of fear a fairly rational strategy.

A complicated cause of crime does not fit easily into a television statement.

Insufficient social services, corruption, poor police work, domestic violence that institutions failed to recognise for years, access to weapons, failures by schools or prosecutors all require documents, time and responsibility from multiple actors.

One “monster” is far simpler to communicate.

And here lies another problem with an obsession with individual revenge: sometimes it suits the system perfectly.

When all attention is focused on how many years in prison one person deserves, there is far less room to ask what preceded the crime and which institution could have acted earlier.

What happens when people do not trust the courts?

It would be far too simple to describe every public demand for immediate punishment as primitive.

Sometimes people rush to issue their own verdict precisely because they do not believe institutions will ever issue one.

In Serbia, that context is not marginal. In its 2026 Rule of Law Report, the European Commission stated that only 36 percent of citizens rated the independence of courts and judges as “fairly good” or “very good”.

That is an increase from 30 percent in 2025, but the Commission still assesses public perception of judicial independence as low.

When people have experience of lengthy proceedings, selective prosecution or political pressure, patience with the phrase “let the institutions do their job” understandably weakens.

The demand for accountability can then be entirely justified.

The public has the right to ask why someone has not been arrested, why an investigation is stalled, why evidence was not secured, who ignored earlier reports and whether political power is protecting someone.

That is exactly what journalists should investigate.

But distrust in a bad system does not make an Instagram comment a more reliable investigation.

It may simply create a second court, even less transparent than the first.

The presumption of innocence also protects people we despise

The presumption of innocence often becomes unpopular precisely when it is most necessary.

It is easy to defend in theory.

It is much harder when there is a video, when police detain someone, when horrific details are published and when an entire society wants punishment.

But legal rights that apply only to sympathetic defendants are of very little value.

The burden of proof exists because police can make mistakes.

A prosecutor can make mistakes.

A witness can lie.

A media outlet can publish false information.

A video can be misinterpreted.

A confession can be problematic.

A first-instance judgment can be overturned.

European rules on the presumption of innocence therefore explicitly require public authorities not to present a person as guilty until guilt has been established according to law, and require the burden of proof to remain with the prosecution.

At the beginning of 2026, the European Court of Human Rights found a violation of the presumption of innocence in Kaya v. Belgium, including because statements made by a prosecutor to the media could lead the public to believe that the person was guilty while proceedings were still ongoing.

That shows how important this boundary is even when the speaker is an institutional representative who knows the case.

On social media, such a boundary barely exists.

Accountability requires more than a culprit

There is another distinction that public debate often loses.

Punishment and accountability are not the same thing.

A person can be convicted without us ever learning why the system allowed violence to continue.

A director can be dismissed while the procedure that enabled the abuse remains untouched.

A police officer can end up in court while the chain of command remains outside the investigation.

One abuser can be imprisoned while ten previous reports made by the victim remain unanswered by institutions.

Revenge is satisfied when someone suffers.

Serious accountability must establish what happened, who did what, what can be proved, who failed to act and how the same mechanism can be prevented from happening again.

That is why facts are slower.

They do not arrive in a single viral clip.

Sometimes they even destroy the story we had already decided to believe.

A society that produces a verdict faster than facts can gain, within hours, the feeling that chaos is under control again.

The next day, much harder questions remain:

Who is actually responsible?

What can be proved?

Why did the system fail when it could still have prevented the harm?

Anger has no answer to those questions.