When the trial ends, but the punishment continues

It's Monday, eight past seven. On the HR director's desk is an extract from the criminal record without a single entry. On the monitor in front of her, a headline that is nine years older lights up, in which the applicant's name is next to the word "accused".

Two minutes later, the interview is no longer taking place. No one will tell the applicant why. He may be legally blameless, acquitted or rehabilitated after serving a sentence. In practice, however, he encountered a different system: with the first search result, a quick judgment and a decision against which there is no appeal period.

The intuitive answer is simple. After all, the Internet only preserved the true message. What once happened can't stay away. In addition, the employer has the right to protect the company and the public has the right to know with whom he has the honor. To delete an old article would mean rewriting history.

But already in this sentence at least four different things were mixed up: historical fact, legal guilt, current the relevance and ease with which a person can be identified by history. The Internet Archive can save in one line. On the contrary, the law must separate them from each other.

First doubt "Case closed" can mean four different things

A person acquitted is not the same as a person who has served a sentence. He is not in the same again situation as a convict whose conviction was later expunged. And neither of these groups is identical to a person whose final judgment was annulled as defective.

Exoneration does not function as a universal historical certificate that no event described ever happened didn't happen It may have several legal reasons. However, it is the binding end of criminal proceedings: public power neither the media cannot continue to treat a person as if his guilt has been conclusively proven. In 2023 the Czech Supreme Court confirmed the compensation of CZK 700,000 to the nurse, which the publishing house in the hundreds of the articles with expressive nicknames giving the impression of a criminal even before the legal one a guilty verdict. The problem was not the very existence of a publicly watched criminal trial. The problem was that the headline got ahead of the verdict.[1]

In the case of an executed punishment, the guilt legally remains, but the formal sanction has ended. The formal consequences continue: Czech the criminal code stipulates that the offender is looked upon as if he had not been convicted after being expunged. Criminal Code at the same time, it says that a expunged conviction must not be reported in the extract from the Criminal Register.[2] It is a legal fiction with real effects — not a physical erasure of every archival reference, judicial a file or a human memory.

Law can therefore change the status of a person. But he can't override the behavior of anyone who types his name into search engines. And that is where the difference between legal rehabilitation and social rehabilitation begins.

At the same time, legal rehabilitation is not a reward for forgetting or a claim that the past has lost its meaning. It is a decision about what consequences the state is still allowed to associate with a past act. At the smoothing the law consciously creates the fiction of non-conviction precisely so that a person is not blocked by everyone after meeting the conditions a formal integrity inquiry. In the case of acquittal, the logic is even more fundamental: if the state does not prove guilt, he cannot leave suspicion to do the same work as judgment.

Social rehabilitation does not guarantee anything like that. It has no official date, form or legal moment power. Some people get it quickly thanks to the support of their surroundings, some never. For a person with a rare name, only one can the old article to act as a lifetime identifier; for a person with a common name, the same message appears results will be lost. The same legal outcome can thus lead to a radically different future according to technical and social circumstances, which the court did not consider at all when imposing the sentence.

Mechanism The internet doesn't just remember. The internet ranks.

The sentence "the internet does not forget" sounds almost scientific, as if it were a property of the material. In fact the internet forgets all the time: links disappear, websites shut down, databases are rewritten and algorithms change order of results. What we perceive as memory is a mixture of archiving, indexing, popularity, optimization, citations and platform decisions.

The difference is substantial. An article can remain in the newspaper archive as a document of its time without at the same time became the first answer to every search for a private person. Preserving a historical text is not the same as connect it permanently to a digital identity. The archive answers the question: "What was written then?" Searching by name answers another: "Who is this person today?"

In addition, the digital environment cancels out the natural friction that memory used to have. A paper cut was necessary find in the archive, know the date and know what one is looking for. Today, a name and a few seconds are enough. Information se did not become truer; only the cost of finding it has dropped dramatically. Change in availability therefore makes a difference and its social effect, even though the content of the article itself remains the same.

Time collapse also occurs. Caption from the year of the charge, the sentencing report and the later acquittal on screen does not rank as chapters of the court file. They compete for a position according to the rules of the search engine. Older, more shared and more emotional text can defeat a more accurate update. The case chronology then turns into a leaderboard attention — and the first place is mistakenly read as the most important truth.

The Court of Justice of the European Union recognized this difference. For old criminal procedure sites that capture an earlier stage and no longer correspond to the current state, the operator of the search engine should assess the seriousness of the matter, result of proceedings, elapsed time, role of a person in public life, current public interest, form of publication and its consequences. And even when the link is not removed, the results should be arranged so that the overall picture reflected the current legal status.[3]

This is the first fundamental change of perspective. The dispute need not be between two extremes — total obliteration and eternal public availability in the strongest possible form. Among them are updating, supplementing the result, reordering, restriction of search by name, deindexing and anonymization. Each tool protects something different and otherwise strongly affects the public memory.

Psychology of information The accusation is the story. The correction is usually a footnote.

The first report offers a plot, conflict and a specific suspect. Later acquittals often come after months or years and tends to take a less dramatic form: appeal court, procedural reason, lack of evidence, canceled statement. The public does not read the court file from the beginning again. It only adds a new detail to the model, which she had already created once.

Psychological research knows the so-called persistent effect of misinformation: it can influence the original information judgment even after it has been corrected. A meta-analysis of thirty-two studies with a total of 6,527 participants found that the corrections, on average, did not completely remove the original effect. They were more successful when they came together from the source of the original claim and offered a comprehensible alternative; were less effective after repetition or with a long time gap.[4]

However, the old report of the accusation may not be disinformation: it may have been accurate at the time of publication actual process status. The problem arises when the reader infers "accused" from the true word false "guilty", or when a long-passed phase without updating is used to describe the present. Therefore, the psychological mechanism does not explain the existence of the archive, but the resistance of the wrong conclusion, which may have formed around him.

This meta-analysis did not specifically examine the reputations of acquitted defendants, and cannot calculate how many one will lose job opportunities because of the old title. However, it offers a plausible mechanism: "the message was later corrected" is not the same as "the original impression disappeared".

Data The second sentence is not imposed by the court. Small decisions shape a person’s return.

Community punishment usually does not take the form of a large public sentence. It arises in parts: unanswered resume, canceled lease, ended acquaintance, workplace gossip, parenting group that "he just doesn't want to take the risk". Each individual person can consider their decision to be a reasonable precaution. However, the sum of these decisions will create a sanction, the length of which has not been determined by any court.

A field experiment in Sweden sent employers paired dummy applications that differed information about previous convictions for assault and community service. Past criminal case reduced the probability of a positive answer by profession by seven to eighteen percentage points. The effect did not appear in all nine occupations examined, which is as important as its existence: stigma is not a uniform automaton but depends on the type of work and the context.[5]

A US study of expungement in Michigan adds another piece to the picture. Within five years of establishment only 6.5 percent of eligible people achieved legal entitlement. For those who have acquired them, wages increased an average of more than 22 percent over their previous year during the following year trajectory; a significant part of the change was the transition from unemployment or very weak work engagement in more stable work.[6]

Even here it is not fair to say that the flattening itself caused all the growth. People who complex process those who successfully complete it may differ in terms of motivation, support and life situation from those who do not apply for it. Moreover, the result comes from a single American state, not from the Czech Republic. But the data is convincing they show two things: a legal claim alone is not enough, and a change in the visibility of the past can be linked with a very specific economic change.

At the same time, the employer does not have to act out of malice. He may feel that he is only reducing the risk: he does not know the applicant, he's short on time, and an old crime story offers him a readily available signal. But such a signal can be outdated, professionally irrelevant or capture proceedings that ended differently. Rational caution on at the level of one decision can then turn into a systematic obstacle at the level of the entire company.

This is where feedback occurs. It is more difficult for a person to get a stable job and housing, therefore he has fewer opportunities demonstrate reliability. The lack of a new credible history makes the old event remain his the most visible public information. The past is not only preserved by the Internet archiving it, but also by limiting the emergence of a competing presence. It is not a sure way to relapse or an excuse for the next one illegal action. It is a mechanism by which reputational stigma can prolong its own conditions duration.

Law and technology "Delete or leave" is an ill-posed question

The General Regulation on the Protection of Personal Data gives the right to erasure under certain conditions, but at the same time contains exceptions, inter alia, for freedom of expression and information, legal obligations, public interest, archiving or legal claims. In addition, data on criminal convictions have a special regime. Czech law allows processing for journalistic purposes if it serves them in a reasonable way, while having also take into account the sensitivity of conviction data.[7]

So there is no button that automatically deletes the past from the media after serving the sentence. Just the same so there is no general rule that anything once published must be forever traceable by full name and without adding the result of the procedure.

Even the popular expression "right to be forgotten" is therefore misleading. It is often not a claim to delete an event from shared history, but about limiting the specific link between a name and a search result. Resourceful an article may remain available by topic, date, or archived; it just stops working as automatic personal business card. The law is not trying to order oblivion here. They are trying to regulate the infrastructure commemoration.

Counter evidence Two European judgments, two opposite results

If there was a simple European rule of "anonymize after years", two significant cases would they had to turn out the same. They didn't work out.

These judgments do not contradict each other. They show that memory is not an absolute value separated from purpose. At one the case's name remained an essential part of the well-known crime and subsequent public debate. For the second, decades later, it did not add to the understanding of the event so much as it took away from human possibilities not be forever defined by one act.

In addition, the form of correction was decisive. The Belgian judgment did not order the erasure of the accident history either prohibit newspapers from keeping an archive. He separated the historical story from the full name of the private person. That's it a smaller, more accurate intervention than disposing of the article.

The second perspective Sometimes the public really needs to remember

Rehabilitating a reputation too easily can harm other people. The victim has no obligation to rewrite his own experience only because the offender carried out the punishment. The newspaper archive is not a private file of convicts; it also documents the work of the police, courts, media and institutions. For public officials, striving people power, regulated professions or activities based on extraordinary trust may be an old criminal case stay legitimately relevant longer.

A complete erasure can prevent the control of repetitive patterns: corruption, financial fraud, professional failure or violence. It may also favor those with money for lawyers and reputation agencies. The "right to a fresh start" would then easily turn into a privilege to buy a cleaner search engine, while for others, their past would remain.

Nor does an exoneration itself mean that every unpleasant fact in the original article was false. A criminal court decides guilt according to precise rules and a high standard of proof. Public interest sometimes also refers to demonstrable circumstances that do not constitute a criminal offense by themselves. Therefore, the medium does not need to be deleted full text only because the prosecution failed. However, they must distinguish exactly what was claimed from what was proven and how the proceedings ended.

So the more accurate question is not, "Does society have the right to remember a person's worst moment?" It reads: "What form should this memory take, what purpose does it serve, to whom is it accessible, how easily and for how long a long time?”

The same information may be appropriate in one relationship and inappropriate in another. Old conviction for financial fraud may have a different weight when choosing a manager of someone else's property than when looking for a storekeeper. Information about a violent crime may or may not be relevant to a particular security-sensitive role justify a public profile available to every neighbor, business partner or new acquaintance. Relevance is not a property of the record itself. It arises only in the relationship between the past act and the present one the purpose and degree of trust required.

Such a distinction is inconvenient. It requires individual judgment instead of a universal prohibition or automatic publication. But discomfort is the price for protecting history at the same time, security, media freedom, victims and the possibility of return. A simple rule would inevitably sacrifice one of these values each time, even if the particular case does not require it.

Current shift The judgment database is not yet a journalist

In July 2026, the Court of Justice of the European Union added an important boundary to this debate. He judged the Swedish a paid online database that made it possible to search for public documents about criminal proceedings. The court admitted that even paid online processing of conviction data can have a journalistic purpose. Not enough however, just open the database and stick a "freedom of information" label on it.

According to the court, it is supposed to be the publication of information, opinions or ideas in accordance with professional ethics rules, after editing or adaptation — or at least according to the actual editorial policy — and after verification factual statements. In the case of simply making available unedited public documents without such work, they do not have to be the journalistic exemption applicable. The final assessment of the specific database remained with Sweden court.[10]

The truth of the document itself did not make the difference. It was created by the purpose and method of processing. Judgment in a forensic database, an investigative article, and a person's commercial profile may contain the same names and the same sentences, but socially and legally fulfill a different function. The public has a strong interest in quality journalism. They may not have the same strong interest in a machine that produces permanent, easily marketable criminal records personal labels.

A more accurate interpretation Society does not need amnesia. It needs adequate memory.

Adequate memory does not begin with erasure. It starts with accuracy. The old indictment article is supposed to visibly contain end of proceedings. The searcher should not build a long-overdue phase above the current legal result. Private the name should remain in the text years later only if it really contributes to public understanding, not only to easily find a person. And complete removal should be reserved for situations where the content is false, unlawful or without an overriding legitimate purpose.

This is not a ready-made algorithm or a universal legal test. It is an editorial model that translates European criteria into comprehensible questions:

• Is the information still accurate and complete, including the final result?
• Is it a public figure or a person who does not return to the public debate?
• Does the full name add to the understanding of the matter, or does it merely enable a reputation check?
• How serious was the event and how much time has passed since it happened?
• Is there a current security, professional or political risk?
• Can the public interest be preserved by a milder instrument than deletion?

The last question is often the most important. An update may be preferable to a deletion. Deindexing preferable to anonymization. Anonymization is preferable to deletion. And sometimes it is correct to leave the name i article without change — for example, because of extraordinary public importance, subsequent public role or compelling current relevance.

However, a procedural issue remains: who is to do this weighing. The editors know the original material, but it can be bound to her old decision. The search engine sees the reach and ranking, but does not judge the evidence as such court. An authority or a court offers independence, but it comes only after a dispute that costs time and money. None of of actors thus lacks a complete context, technical power and democratic legitimacy at the same time.

A reasonable system therefore does not need a single all-powerful arbiter, but correctable steps: available update request, reasoned de-indexing decision, possibility of review and transparent criteria. Otherwise, there is a double risk. Either the public memory is secretly decided by a private platform with no control, or any remedy will become so expensive that only people with legal and financial backgrounds will obtain it.

This is where the original dispute changes. It's no longer just about privacy versus freedom of speech or the wrongdoer versus companies. It's about the separation of powers. The court may impose a penalty within the legal limits. Search engine, employer, neither neighbor nor anonymous databases have such limits. When the social sanction arises from the sum of their decision, may be longer than the punishment that the state considered reasonable — and may also fall on a person whose guilt the state ultimately did not prove.