Can AI Be The “Third Umpire” In The Judicial System?

There is little doubt that a majority of Indians — perhaps the vast majority — are convinced that our judicial system is in need of a root and branch overhaul.

We have all heard multiple horrible personal anecdotes of a justice delivery system in disarray, to put it as mildly as possible.

There are mind-boggling statistics of what really represents a deep failure of this system. A visit to the online National Judicial Data Grid (NJDG) shows you that there are 28 million pending cases that are more than a year old, out of 44 million cases in total currently being handled. 

Few will deny that the problem needs to be tackled with some radical ideas, and that technology will have to form the base of any solution. And there is precedent for it. 

Who would have thought in 2013 that every Indian would have a digital personal identity, access to bank accounts, direct delivery of state benefits to individual recipients, and would be so financially enabled that we barely need to carry cards, let alone cash, anymore?

Why should this not be possible for the delivery of justice, something every citizen has the right to expect in the New India that a majority hope to see?

We can leverage technology for this purpose — as we have done elsewhere with the Digital India initiative — deploying Machine Learning and Artificial Intelligence. There is no shortage of top-tier expertise, both at the managerial and the technological levels.

From a data point of view this should not be too difficult either, as a lot of the information about past cases, rulings as well as laws, amendments and bills should be in digitised form already. 

If an AI-based justice delivery system can be created simulating the role of judges at various lower levels of the judiciary, — templated case data input should generate a maximally impartial outcome that can be relied upon by all parties. 

Sure, critics will say that there will algorithmic bias depending on who is doing the programming, etc. But equally surely, the bias itself will be far less than in the case of any human judge? Subjectivity levels are certain to be far lower.

Of course, no one is suggesting that the entire judiciary should be replaced by AI. Far from it. The objective is to have a support system that can dispense with the (often routine and procedural) issues that clog the system. 

In fact there are many ways in which the system can be rendered far more efficient than it currently is, and yet leave the human factor available to the litigants.  

It would appear that the first thing to do would be to apply such a system to the millions of pending cases. Once the system is up and running, using our own supercomputer capabilities, this should not take more than a year at most — at least in terms of pending case disposal. 

A huge amount of commercial activity would be generated to execute sentencing outcomes. Finally, it would be another example of India eliminating corruption-derived bottlenecks as well as revolutioniisng the legal system through a digital initiative. 

Possibly another first in the world that we as Indians can be truly proud of. 

Nothing is fail-proof. Just like with AADHAR/PAN/GST there will be exceptions to the rule, and we will not get it right immediately from kick-off. But as the machine learns and the AI engine fine tunes itself, there will come a point where human intervention requirement will be reduced to almost zero. 

Would such a system be perfect? No. But that is not the correct question. 

The correct question is: Would such a system be far less imperfect than the arbitrariness and chaos that ordinary people are forced to live with? Will it make our democracy more trust-driven, resilient and reliable?

The answer to both questions can only be yes.

There is also a psychological advantage that comes with AI. Owing to a millennium of foreign rule, many (if not most) of us Indians have a psychological predilection which can be called the "foreign coach syndrome". 

Even in cricket, where we are and have been champs for a long time, we used to seek the leadership and guidance of foreign coaches (equivalent in some sense to the foreign rulers of the past). It was assumed they would be “more fair” than we ourselves.

Well, the AI judge negates this "foreigner coach" bug in our minds, because it will be truly neutral. Even better, it will be non-human. The thing is, we are used to this too. 

Using the cricket analogy again, note that when the human referee fails, we always go to the “third umpire”. AI can be the “third umpire” for our judicial system.