Predictive Justice: Using AI for Justice Introduction

By Bhishm Khanna
Outcome prediction in a legal case has always been crucial to the practice of law. Predictive analytics using AI tools have made predictive justice possible and offer sophisticated means for outcome prediction of disputes that are presented before the courts of law. Predictive justice refers to using the analysis of large amounts of data by means of AI-enabled technologies for predicting outcomes of legal disputes.

AI models allowing predictive analytics can offer solutions to a number of problems faced by our judicial systems such as long pendency, legal inconsistency in application of law, poor triage of disputes etc. As evidenced from various research studies, the accuracy of AI predictive models is more than the lawyers with expertise in the relevant area of law.

Predictive models have the capability to exclude legally irrelevant factors which would help in standardisation and elimination of application of arbitrary factors in judicial making, thereby allowing legal certainty. Vast deviance of the conclusion arrived at by judges from the AI could be indicative of reliance on these arbitrary factors.

A dispute filed before a court of law where violation seems to be likely (as concluded by predictive AI) could be prioritised. Cases that involve simple application of law can be automated. Judges can, thus focus on these cases and increase the efficiency of the court system. Parties to a dispute may also decide to take the case outside the court in situations where success is unlikely (as predicted by AI) and not waste time and resources in long litigations. This would help reduce the information asymmetry between the parties. Use of AI in judiciary can also instil public trust and thereby can help deployment of AI in other public delivery systems.

Despite the advantages, there are valid concerns in terms of jeopardising independence of the judiciary by AI which is often voiced by stating that AI will make the judges redundant. However this is not the case. Use of AI, it is suggested, must be accompanied by human oversight and supervision. Its output must not be considered conclusive. Machine based decisions must only assist judges and not replace them.

This can also help in addressing the issues of over reliance on the conclusion of the machine, called the automation bias or algorithmic bias where poor dataset results in inequitable conclusion of the machine etc. This oversight is not just by the judicial officers but by other stakeholders as well, so that explainable AI can be developed. An explainable AI would help in achieving accountability by design. It would also allow for legitimate challenges to the conclusion reached by an AI.

Research in public universities in collaboration with the private sector can be beneficial in building strong predictive AI models and therefore must be leveraged. Development of tools for predictive justice would necessarily require a contribution from various Ministries, civil society, think tanks, research centres etc. Only a robust consultation and regular audit could ensure that the common citizens are not further alienated from the judicial process and its development or functioning is not influenced by private actors seeking profits, thereby maintaining the sanctity of the justice system.

Putting in place predictive technology allows for an automated system that can resolve various issues plaguing our courts without their constant intervention thereby achieving the ideal of minimum government and maximum governance. AI is bound to cause disruption in the legal sector; therefore, a pragmatic approach would be to take advantage of the same. AI in courts would complement the vision of NITI Aayog of #AIforAll and the other initiatives taken by the judiciary such as SUPAC and SUVAS. Better legal contextualisation and capacity building demands starting the process early to plug in the loopholes in time. This could aid the AtmaNirbhar Bharat Mission aswell. This requires a framework of action in the form of a strong policy that defines the scope of predictive justice in courts of law and puts in place necessary safeguards.

This article was written by Bhishm Khanna. Bhishm Khanna secured first prize at the Atlas- CPPR South Asia Public Policy Challenge 2020- 21.


more recommended stories