Legal Articles

Artificial Intelligence (AI) and Legal Institutions

Written by Aruneshwar Gupta

In 1972 Bobby Fisher took over title from Boris Spassky to become the World Champion and in the same year he played three games against the Greenblatt Program and massacred the computer. In his comments on the games, Bobby said that the computer was very weak and that he could give it great odds and still beat it. In 1989 IBM took up the challenge and build up a computer chip ‘Deep Thought’, which evolved to ‘Deep Blue’, ‘Deep Blue Jr.’, ‘Big Blue’ and finally ‘Deep Thought II’ as it continued to play against grand masters over and over again.

In 1997 Garry Kasparov was hunched over a chessboard, visibly frustrated. He was fidgeting in between turns and shaking his head in disbelief as he waited for his opponent to put the final touches on an inevitable victory. Finally, Kasparov made his move, stood up and raced away from the board.      He raised his arms, astounded that he was beaten by a machine. His opponent was the IBM supercomputer, a machine that was capable of imagining an average of 200 million positions per second. Going into the match Kasparov was confident, he was the Michael Jordan of chess. He had been beating chess-playing computers and was considered nearly unbeatable.

When Kasparov, one of the greatest chess players of all times, lost to a computer in front of a global audience, Artificial Intelligence was out and open and the time had begun when machines were ready to surpass humans in processing data and accurately displaying results, based thereon. The computer continued to use the logic of the grand-masters in arriving at the result and using the final result as new information. By that process the computer reached at a more refined inference and was able to infer at much quicker pace, though it lacked originality or a new idea with which it could experiment, which faculty is unique to human beings.

The real purpose of AI was and will always be to help homo sapiens to evolve further as better human beings. The direction of human evolution and all institutions evolved by him including law, is to create a harmonized, conflict free civil society without exploitation and corruption and to provide compensation where ever there is misuse of power, will has been suppressed, injury has been caused and damage has been done.

A chess board has 64 squares, 32 black and 32 white and has 32 players on the board, 16 black and 16 white. Move of every player on the board is defined by rules of the games. The entire game is about placing of each player strategically to checkmate the opponents’ king. With every move the position of the board changes and that leads to making a new move by the opponent. Over the years, 1972 to 1997, the computer continued to absorb the moves of the grand-masters with whom it played and processed the next best move based on the logic of the location of the players on the chess board, by the grand-masters. The next best placement evolved by looking at the moved played by the opponent, the position of the entire board i.e. the placement of all the players on the board on black and white squares and the possible and likely moves that may be made in future. The next move was the result derived from the sum total of all the positions and moves played by all the grand-masters against it all these years.

In every litigation there are a certain set of facts existing prior to the filing of Suit or FIR. There are about 1250 Central Acts and about the same number in each State and Union Territory, more are enacted and amendments are brought about to the existing enactments regularly. There are Rules, Regulations, Notifications, Memorandums and Government Orders (‘Statutes’) existing and issued almost every day. Supreme Court has in it’s about 50,500 judgments (1950 till date) considered more than 4,000 of such statues and other notifications etc. having force of law. There are about 10 lakh judgments delivered by High Court all over the country and have interpreted a large number of statues and notifications etc. as and when any issue arose before it.

This is like the setting of a chess board. The parties and their advocates strategically apply statutes and judgments to the facts of their case and by a process of dialectics make all out efforts to get justice in their favour. Unlike chess where the winner is decided by checkmating the opponents’ king, in litigation the Judge decides the winner as he reaches to truth and justice by a process of logic and reasoning based on his perception and experience. In every case the judge has to give grounds for his judgment and hold one party as winner and other as looser on the basis of the information, understanding and appreciation of the facts stated in documents and pleadings, relevant statutes and judgments and his insight into socio economic milieu. He has to compensate the person who has been wronged in civil cases and punish the guilt in criminal cases.

The documents and pleadings in every case are unique to itself. No two faces are same and no two cases are same, they may be similar or even twins. That makes law different from chess. The society is dynamic and evolving and so are statutes and abrasions. Statues are enacted to deal with abrasions or to regulate the social relationship to restrain abuse, exploitation and corruption. When Kasparov was beaten the computers’ were capable of making 200 million calculation per second and had limited storage capacity. Today they can make up to 10,000 billion calculations per second and have unlimited storage capacity, everything is on the cloud.

Application of AI to legal institutions is irreversible and unalterable process of evolution and we will have to adopt the same sooner or later. The sooner we accept the change, earlier we will be able to evolve further. It is the most meaningful and phenomenal change for which we have been waiting for a long time, to wipe out every tear from every eye, by bringing justice to all. Once an authentic data on uniform parameters of all the judgments of Supreme Court and High Court, relevant Statutes, pleading and documents therein is created, AI application can provide the following help to all the stake holders, which will continue to become more meaningful and significant as it will be  made more and more applicable:

  1. Benefit to litigants:
    1. It will be possible to read and cull out relevant data from any number of documents and sorting them on the basis of dates and also on relevance. Once the appropriate data is extracted out of the documents, the facts and events can be arranged in strict chronology. On the basis of the information thus extracted, further help can received in drawing out a list of dates and events. In most cases the facts and events are not disputed, in Singapore the advocates have to draw a common list of dates and events. What is in dispute is the understanding, interpretation and the reason behind the event.
    2. Afteranalysing the dates and events and the existing database of judgments and statutes, the application can display the breaches committed by the parties and the rights of the parties which have been violated.
    3. Once the terms and conditions of contract that have been breached are determined or the provisions of statues that have been violated becomes available, the application can provide the list of judgments that will support the case.
    4. Finally it will also be possible to determine
      1. the possible time consumed in reaching to the end of litigation
      2. the cost of litigation
      3. whether it will be meaningful to take up the litigation or not
  1. Benefits to Advocates:
    1. Drafting of Deeds, Wills, MoU, Agreements are already in vogue. But it is difficult to lay your hands on formats which are tailor made for the purpose. AI application can help in reaching to a more authenticated format to be used.
    2. Authenticated translation of documents from local languages to English, particularly at the High Court and Supreme Court level still continues to be serious issue. As the database of documents, pleadings, submissions and judgments continue to grow in both vernacular and English Language, more authenticated translations can be made available at a much quicker pace.
    3. Use of AI application will certainly improve the quality of pleadings by eliminating grammatical mistakes and using legal language, spell check is already available. Once a true and correct list of dates and events and pleadings is ready, the next step will be to help in preparing:
      1. Issues
      2. Question of law
      3. Grounds in support of the pleadings
    4. Once the pleadings are completed AI application can help in matching them with the oral evidence that come on record.
    5. After the pleading and evidence is completed, the application can help in framing of submissions.
    6. Finally the application can provide for relevant judgments in support and against the submissions.
    7. Based on the entire litigation material the application can also evaluates the pros and cons of the submissions and then assigns a score for each submission.
    8. AI can radically make legal research much easier. The humongousredundant results by word search and search within search can be minimized. Advocates always need precise and only the most relevant judgments to be made applicable to the case in hand as precedents. Judgments can be recommended for the specific facts and submissions of the case, which fits their need.
    9. Once the parameters for appointment of senior advocates and judges of High Court are objectively defined, it will be possible for those who wish to make that stream, as their life objective to follow the parameters and reach his/ her avowed ambition for which s/he is willing to work with humility, determination and perseverance.
    10. At the time of elevation of an advocate as judge, it will be possible for her/ him to follow well defined path and be aware of the others in the field, whom s/he will face for appointment as Chief Justice of High Court, judge of Supreme Court of India and Chief Justice of India. A substantial amount of prejudices and preferences will be effectively eliminated.
  1. Benefits to Judges (at all level):
    1. Make the understanding of facts simpler and easier.
    2. Provide options in analysingthe facts and decision making process
    3. Provide options for granting punishment to accused and compensation to victim in criminal cases;
    4. Provide options for granting of relief, compensation, damages and costs in civil cases
    5. Help in usage of better language and usage of right words and terms
    6. Improve writing of judgmentsby providing the ground and judgments which supports the final view taken
    7. Identify tonal nuances of the judgement being written and help making them neutral and objective.
    8. Bring about content quality and  consistency
  1. Benefit in appointment of High Court Judges, Chief Justice of High Court, Judges to Supreme Court and Chief Justice of India:
    1. Once the data on the performance of each advocate in each case from the inception of her/his career, becomes available on objective parameters, it will be possible to pick up the best talent from the bar for appointment as judge of High Court. AI will be able to retrieve the following data for each advocate:
      1. Petitions and Counters drafted – Subject and issue wise
      2. Admission matters argued – subject and issue wise, including time spent on legs;
      3. Final hearing matters argued – subject and issue wise, including time spent on legs;
      4. Written Submissions submitted in court or Propositions of law propounded;
      5. Articles and books published;
      6. Presentations and Addresses delivered in national  and international conferences
      7. Judges can continue to enter and mark conduct and performance of each advocate in case in which they consider it noteworthy on the parameters like presentation, dignity while addressing, pitch and tone of voice, mannerism, gestures etc.
    2. Once a person is appointed as a Judge of the High Court, report on her/his performance can be objectively retrieved from the proceedings, orders and judgments delivered. Analysis of every Judge can be made by using multivariate regression method of her/his as a single puisne judge, puisne judge along with a senior judge and as presiding puisne judge. Some of the parameters can be:
      1. Proceedings issued
      2. Orders delivered – subject and issue wise, including time taken
      3. Judgments delivered – subject and issue wise
      4. Appeals and Special Leave Petitions filed against orders and judgments delivered
      5. Orders and judgmentsconfirmed and set aside by Supreme Court
      6. Judgments of Supreme Court, High Court and other courts referred, considered, analysed, explained, applied, distinguished and held obiter etc. in the judgments delivered
      7. Ratio between proceedings, orders, signatory and judgments delivered
    3. Propositions of law propounded in judgments
    4. An algorithms can be developed to use the above data and identify the perceptions of the judge and her/his proficiency in bringing justice to the litigant. It will also be possible to identify inconsistencies in judgmentsin respect to a particular issue arising in a subject. This data can be used meaningfully for the purposes of her/his elevation as Chief Justice of the High Court and elevation as Judge of the Supreme Court.
  1. Benefit to Legislature and Executive:
    1. There are three basic reasons for a litigation:
      1. The statute is invalid,
      2. The statute is valid but the action is invalid,
      3. The person has taken some advantage by violating a provision of statue or breached a term and condition and does not wish to compensate back to the state or the person from whom the benefit have been taken by him.
    2. Normally 50% of litigation is civil and 50 % is criminal and in civil cases 85% are those in which Government is a party. Out of the 3.3 Crore pending  litigation, an AI application can be run on the about 1.40 Crore Civil litigation in which Government is a party, to determine under which of the three heads the case is falling.
    3. If the Legislature understands and realizes that the litigants have approached the court because the statute is invalid in view of law laid down by Supreme Court, they can make amendments to the Statute and that will lead to disposal of all those cases. If the Legislature still persists on continuing with the litigation, the Court must burden the legislature with heavy costs;
    4. If the Executive understands and realizes that the litigants have approached the court because the statute is valid, but the action taken by them is invalid, the executive can retrieve the order and get the cases disposed of. If the executive still persists on continuing with the litigation, the Court must burden the Executive with heavy costs;
    5. If the analysis is that the person has taken some advantage by violating a provision of statue or breached a term and condition of the contract and does not wish to compensate back to the state or the person from whom the benefit has been received and is using the process of court to stall such recovery. The court should give him an option to withdraw the litigation and pay back the advantage received by him. If the litigant still persists on continuing with the litigation, the Court must burden with heavy costs;
    6. A mature AI application can over the years help the legislature and executive in ascertaining the validity of a statue or action at the time of its enactment or passing of the order and the possible impact it will make on the exiting judgments and statutes.

All this can dramatically reduce the mechanical work of the Advocates and Judges and make their contribution to justice more meaningful and purposeful. They can take time in thinking more than what is already made available by the machine. However, AI will never ever replace the ingenuity of an advocate and experience of a Judge and that will be the real test of her/ his value worth as a lawman, to think out of box which is working with her/ him shoulder to should.

The final vision of application of AI to Legal Institutions is that before entering into litigation, the person should be aware of his coming karmas. Before filing the case and during the process advocates are able to put their best foot forward. Before passing an order or delivering a judgment, the same could be thoroughly examined. Finally as soon as a statute is enacted or a judgment is delivered, it must be possible to objectively say “This is Law! This is Justice!” or “Is this a Law? Is this Justice?” This will be the true application of Artificial Intelligence to the legal institutions, not just creating some meta-tags to help retrieval of judgments on subject and issue, which application has now become ancient.

Aruneshwar Gupta
Senior Advocate, Supreme Court

About the author

Aruneshwar Gupta

Leave a Comment