Balancing rehabilitation and public safety under AI-powered governance: a three-stage mixed-methods empirical analysis of misdemeanour record sealing in China

In the modern world, a criminal record is more than just a court verdict; it is a permanent digital shadow. In many places, once a person is convicted of a crime, that mark follows them for life, closing doors to jobs, housing, and education, often affecting their family members as well. This is especially true for minor offenses, where the punishment of a few months in jail or a small fine is followed by a lifetime of exclusion. As artificial intelligence becomes deeply woven into how governments manage public safety, these digital records are easier to find and harder to ignore. This creates a difficult tension: society needs to stay safe from dangerous people, but it also needs a way for those who have made small mistakes to move forward and become productive members of the community. The question of where to draw the line—deciding which minor crimes should be sealed so the record disappears, and which must remain visible for public safety—has long been a subject of debate among scholars and lawmakers.

A new study from China tackles this question by moving beyond theory and looking directly at the data. The researchers wanted to find a practical, evidence-based way to define the scope of a system that seals the records of people who commit minor offenses. Instead of relying on logical arguments or isolated examples, they combined interviews with legal professionals across ten different regions with a detailed analysis of hundreds of real court cases. Their goal was to measure exactly how many people would be affected by different rules and to identify which specific types of crimes and offenders should be excluded from this protection to ensure public safety.

The study began by listening to the people on the front lines: police officers, prosecutors, judges, and community correction workers. These practitioners were asked what they believed should be included in a record-sealing system. Their answers pointed to three main factors that must be considered: how long the sentence was, what specific crime was committed, and how likely the person was to commit another crime. The interviews revealed a strong consensus that while most minor offenses should be eligible for sealing, certain categories must be off-limits. These excluded categories include crimes that endanger national security, serious violent acts, terrorist activities, gang-related offenses, and sexual crimes against children. The professionals also agreed that people who have a history of reoffending, such as repeat offenders or those with a pattern of habitual criminal behavior, should not have their records sealed, regardless of the sentence they received for a specific minor crime.

To test these ideas, the researchers gathered real court data from two different counties in northern China, covering hundreds of individuals who had been convicted of crimes. They applied six different sets of rules to this data to see how many people would qualify under each scenario. One set of rules suggested sealing records for anyone sentenced to three years or less, while another suggested a stricter limit of one year or less. They also tested a rule that treated intentional crimes differently from accidental ones, allowing a longer sentence for accidental crimes before the record would remain open. The analysis showed that the choice of sentence length makes a huge difference. If the rule allowed for sentences of up to three years, the system would cover a vast majority of offenders, sealing the records of about 81 percent of the people in their sample. However, if the rule was tightened to sentences of one year or less, the coverage dropped significantly, protecting only about 62 percent of the group.

The researchers found that the most balanced approach, which protects public safety while offering a genuine chance for rehabilitation, is to set the limit at public surveillance, criminal detention, or a fixed prison term of up to one year. For crimes that were committed by accident rather than on purpose, they suggested this limit could be extended to three years, though they noted that their data on this specific group was small. Crucially, the study confirmed that specific charges must be excluded based on the consensus of practitioners gathered during the interviews. The qualitative findings identified crimes like robbery, child indecency, and drug trafficking as categories that should be excluded from sealing, even if they resulted in short sentences, due to their perceived danger. Furthermore, the study highlighted that risk is the deciding factor for many. Individuals identified as habitual offenders or those with a high risk of reoffending were systematically removed from the eligible pool, ensuring that the system does not protect those who pose a continued threat to the community.

When the researchers compared the results from the two different regions, they found the findings were remarkably consistent. The proportion of people who qualified for record sealing was similar in both areas, suggesting that these rules could work across different parts of the country. The study also noted that as the number of minor crimes in society continues to rise, the need for a clear, fair system becomes more urgent. Without such a system, the collateral damage of a criminal record—losing a job, being unable to find housing, and the stigma passed down to family members—can create a cycle of poverty and instability that harms society as a whole.

While the study offers a clear path forward, the author is careful to note that their work is a starting point. The data came from a limited number of regions and a relatively small number of cases, so the exact percentages might shift if a larger, nationwide sample were used. Additionally, the rules for identifying high-risk offenders relied on existing records rather than new, long-term tracking of whether these people actually reoffended. Despite these limitations, the research provides the first solid, empirical evidence for how to build a record-sealing system in China. It suggests that by combining a strict limit on sentence length, a clear list of excluded crimes, and a focus on the risk of reoffending, the legal system can strike a balance. This approach allows the vast majority of people who make minor mistakes to rebuild their lives while keeping the most dangerous individuals and the most serious crimes visible to protect the public.