IIIT Hyderabad Study Takes A Data-Driven Look At Fairness In Justice Delivery System

Subjectivity is an inherent part of decision-making processes, whether it involves a doctor’s diagnosis, a manager’s approval, or a judge’s ruling. Various factors such as experience, context, workload, and subtle cognitive biases shape these decisions. While this variation is a natural human trait, it can lead to troubling inconsistencies, particularly in systems that are designed to be fair and objective. A research paper titled “Data Cube for Exploring Anomalies in Justice Delivery: An Experiment on Indian Judgements”, authored by researchers from IIIT Hyderabad, including Sriharshitha Bondugula, Prof. Krishna Reddy P, and Narendra Babu Unnam, in collaboration with Prof. Santhy KVK from NALSAR, reveals that legal systems are not exempt from these discrepancies.

Understanding Judicial Disparities

The researchers argue that judges, like all decision-makers, bring their unique interplay of personal experiences and cognitive processes to the bench. This makes some degree of variation in judicial decisions inevitable. Decisions regarding custody arrangements, bail grants, and sentence impositions are often subject to the judge’s discretion. Such discretion can lead to disparities in similar cases across different individuals and courts, potentially affecting liberty, justice, and public trust.

Identifying Anomalies

Instead of focusing on individual intent or motivation, the research proposes a framework to systematically identify anomalies and disparities in judicial decisions. This approach utilizes data to highlight unexpected divergences in decisions, similar to practices already employed in fields like medicine and marketing. The researchers believe that applying a similar analytical approach to the legal domain can assist institutions in detecting inconsistencies or unfairness in judicial decisions.

The Wide Space for Interpretation in Indian Criminal Law

Indian criminal law provides clear definitions of permissible punishments but allows for a broad interpretation. Under Section 53 of the Indian Penal Code, punishments range from fines to death sentences. While minimum and maximum penalties are specified, the gap between them often leaves room for judges to interpret and quantify the severity of a crime based on contextual nuances. Judges frequently consider subjective factors including the accused’s background, prior criminal history, intent, brutality of the act, the victim’s age, and other mitigating or aggravating circumstances. What may be deemed serious in one courtroom could be viewed differently in another.

Methodology of the Study

To systematically study these differences, the researchers employed Online Analytical Processing (OLAP), a technique commonly used in sales, healthcare, and marketing to analyze large, multi-dimensional datasets. The researchers noted that no previous efforts had been made to extend this data cube-based framework to explore anomalies in the legal domain.

Using the proposed data cube approach, the researchers analyzed sentencing patterns across multiple dimensions, including:

• Crime type
• State
• Year
• Punishment type
• Prison duration
• Fine amount

This multi-dimensional analysis enabled the researchers to zoom in and out of the data to identify trends and outliers.

Transforming Judgments into Structured Data

One of the primary challenges faced by the IIIT team was the extraction of data, as court judgments are not formatted as neat datasets. The researchers manually extracted data from nearly 3,500 Indian criminal cases involving serious offenses such as murder, kidnapping, and rape, covering judgments from 2005 to 2010. They utilized large language models (LLMs) to convert messy, unstructured legal text into structured variables, thereby creating a clean dataset that captured verdicts, punishment types, prison durations, and fines. This structured dataset served as the foundation for their analytical framework.

Insights from the Data

Once the data cube was established, the researchers uncovered notable patterns. The study revealed significant variations in fines and prison terms for similar crimes across different states. For instance:

• In homicide cases, fines were rarely imposed in most states, except for Kerala, which exhibited significantly higher fine amounts and greater variation.
• In aggravated rape cases, Himachal Pradesh imposed higher fines compared to Delhi, despite similar average values in other regions.
• For aggravated kidnapping, longer prison terms were observed in Rajasthan than in Haryana.

These patterns highlighted the need for deeper exploration of these disparities, as they are not readily apparent without a multi-dimensional analysis.

The Balance Between Discretion and Consistency

The researchers emphasize that while discretion is essential for individualized justice, maintaining a balance between flexibility and consistency is crucial for fostering fairness. They cite examples from other countries, such as the United States, which introduced sentencing guidelines in the late 1980s to reduce subjectivity and enhance uniformity. However, debates around the tension between rigidity and fairness continue.

Future Directions

The research aims to encourage the exploration of a comprehensive data cube-based framework to investigate disparities in the legal domain, both in India and internationally. Periodic assessments using such a system could identify disparities and promote corrective measures. Moreover, the researchers have made their dataset, the Indian Judgements Punishment Data (IJPD), publicly available. Future plans include extracting data directly from full judgment texts and incorporating additional dimensions such as the age of the accused, the victim’s age, and even the gender of the judge.

Conclusion

In a justice system that relies heavily on human judgment, this study serves as a powerful reminder that while data cannot replace discretion, it can illuminate areas where discretion may lead to disparity. The findings underscore the importance of using data-driven approaches to enhance fairness and consistency in judicial decisions.

Note: This article is based on research conducted by IIIT Hyderabad and aims to provide insights into the fairness of the justice delivery system in India.