Within India today, AI’s exploitation is arguably linked to improving administrative efficiency and improving decision-making processes in the justice system. For the former, developing task-specific narrow AI tools should be the first generation of AI innovation. These should potentially ease the general rigour of the registry, and also aid judges in spending lesser time on administrative responsibilities, in lieu of judicial work. Further, these will provide sophisticated automation for banal and time-consuming admin processes.
With respect to the latter, the spectrum of possible AI can include tools for intelligent analytics and research, and even computational tools (and predictive justice in the longer run). These tools can provide comprehensive legal briefs on cases, encapsulating pertinent legal research, identifying crucial points of law and facts, and thereby expediting the judicial process. This can effectively supplement human judgment in adjudication. Furthermore, intelligent tools, like legal bots, can be designed to help potential litigants with better informed decision making concerning their legal rights, and easily and cost-effectively access basic legal services.
Before venturing into the proposed use cases, it is pertinent to address a vital prerequisite for any prospective AI innovation for the justice system. Current ML (machine learning) and deep-learning techniques are heavily reliant on accessible data. Once such datasets are readily available, AI-driven technologies can be realised for augmenting administrative efficiency and the quality of decision-making.
Need for open access to judicial data: In India, the judiciary, as well as the larger justice system, openly accessible datasets are a far cry. For instance, judgments, which are public resources, are often not published in machine readable formats. This leads to there being technical hurdles in accessing basic legal databases. A recent study has identified how an open judiciary could spur some innovative tech solutions, including those driven by AI algorithms. To harness the transformative potential of emerging technologies like AI for our justice system, it is critical for it to recognise the impediments in current data access, sharing and usability, and remedy them swiftly. Openly accessible and machine-readable data is a sine qua non for the Indian justice system’s digital transformation. The judiciary should create an open-access policy setting out what kind of non-sensitive data is to be made openly accessible, and laying out some broad rules to govern such data sharing. This is discussed in greater detail in the roadmap section of this paper.
Improving Administrative Efficiency
Process re-engineering and automation: The pandemic has led to a surge in discussion around increasing digitisation through the eCourts Project, creation of virtual courts, and the potential of online dispute resolution. Within this conversation, AI has also become an increasing talking point. For instance, administrative efficiency can be improved by using intelligent and sophisticated innovation to automate daily processes of the registry. For this, task-specific, narrowly tailored algorithms, trained through ML, can be deployed to aid in run of the mill administrative functions, from something as simple as scheduling hearings and creating causelists, to more complex tasks like discovery and review of evidentiary documents. Other small tasks which can benefit with the use of AI include interventions at the level of smart e-filing, intelligent filtering/prioritization of cases or notifications and tracking of cases.
In India, the preliminary work in the use of AI has already commenced. SUVAAS was the pioneer of such task specific algorithms, designed by the Supreme Court’s AI Committee. It relies on natural language processing (an ML process), easing and expediting translation of judicial orders and rulings. Additionally, as was announced last year, the SC AI Committee is also working on a composite new tool named SUPACE (Supreme Court Portal for Assistance in Court Efficiency), which will target different processes like data mining, legal research, projecting case progress, etc. There is also an in-house software being piloted in the 17 benches of the Supreme Court to make them paperless.
While these pilots are promising, there is a need to identify steps for scaling these technologies and their adoption, which will be discussed in the last section of this paper. It is also pertinent to reiterate that the success of these pilots and further innovation, will be contingent on the availability of adequate training data corpuses, and capacity building of stakeholders through training and skill development.
Tools for intelligent legal analytics and research: A significant amount of work for a lawyer or a judge involves legal research, analyses of factual propositions, determination of appropriate legal provisions and other similar mechanical skills. In the Indian justice system, any aid to save judicial or legal time by expediting these processes will be a significant value addition. To enable this, ML algorithms can be conceptualised, designed and deployed for intelligent analytics and research work. Such usage is also witnessing deliberation and adoption in other jurisdictions where AI-driven tech is being integrated for aiding judicial decision-making processes.
Beyond lawyers and judges, legal analytics and research tools that are commonly accessible by the public can improve its engagement and understanding of the law. This engagement is vital for creating a better-informed citizenry, which is more proactive and educated about its legal rights and obligations. Such algorithms can be modelled into tools which can offer preliminary legal analyses, relevant case law, and basic legal advice to potential litigants. For instance, a person may be a victim of a cheque bouncing case and require some basic inputs on how to proceed legally. A bot could present interactive toolkits, prescribing next steps, including identifying facts for issuance of a legal notice, filing FIR, and even provide a prediction of success based on facts and established law.
Computational tools for justice delivery: In India, we have already experienced some amount of process automation in traffic challan cases through the establishment of online payment mechanisms. In addition to expediting judicial processes through process automation, ML algorithms or deep-learning technologies can be used in more sophisticated ways such as developing tools which can help judges in arriving at decisions. For instance, motor vehicle compensation claims are largely calculated based on established principles and variables. The role of the claims’ tribunal is limited and rarely involves legal interpretation. A possible tool could aid the judge in cataloguing the requisite documents for such a claim, and glean the relevant information that will allow the judge to determine if compensation is due, the party that is liable to pay, and the value of compensation. Assimilating learnings from such first generation of computational tools will be necessary before scaling them up for more controversial legal areas such as predictive criminal justice. While predictive justice may sound futuristic, there is growing evidence of its use within justice systems. As Chief Justice Roberts of the US Supreme Court has famously said in response to AI assisting “even judicial decision making, “…it’s a day that’s here and it’s putting a significant strain on how the judiciary goes about doing things”.
That said, it is crucial to ensure the use of computational predictive justice is not devoid of human intervention. The idea underlying it must be to supplement, not supplant human actors like judges. Additionally, predictive justice experiments have also raised some legitimate concerns around perpetuation of societal biases, lack of transparency of decision making, curtailing judicial autonomy, etc. Ensuring that these issues are pre-emptively addressed will be key in adopting these efficiency increasing judicial decision-making support systems. Some of the issues that need to be addressed are discussed at greater length in the challenges section of this paper.
Legal robotics for improving access to justice: AI-designed bots are becoming increasingly ubiquitous across different sectors like insurance, banking, e-commerce, etc. Bots are convenient and interactive tools for providing common information to a user in a conversational format. With respect to the justice system, legal robotics can play a crucial role in serving as intelligent and dynamic repositories of FAQs, which aids the public’s understanding of laws. This would be extremely useful for common citizens and potential litigants in getting basic inputs on a prospective legal case, and making better informed decisions, inter alia whether litigation is needed or not.
In addition to providing better information, legal robotics can also improve access to legal services. For a common person, accessing these, or even grappling with a potential legal situation can be a daunting conundrum.
Intelligent algorithms (or bots) can be useful in furnishing basic legal information to potential litigants and readily connecting them with legal aid services or pro-bono lawyers. Basic legal services like drafting and conveyancing, legal analyses and interactive breakdown on laws, etc., can be some modes for mainstreaming access to such services, without the trouble of locating and paying for expensive lawyers.This is part-1 of a two-part series on AI and judiciary wherein excerpts from the paper ‘Responsible Artificial Intelligence for the Indian Justice System’ are published with the permission of Vidhi Centre for Legal Policy. You can read the entire paper here