Scribd Logo
Upload

Welcome to Scribd!

  • 98 views

    Artificial Intelligence in LAW

    Uploaded by

    Madhu
    This document discusses using artificial intelligence and machine learning techniques to help address problems in India's legal system. It outlines issues like a large backlog of pending cases and lack of judicial resources. It then suggests how AI could help with legal tasks like research, contract analysis, and verdict prediction by analyzing past cases and contracts to find patterns and predict outcomes. It proposes an algorithm using long short-term memory to train a model on previous judgments and then validate and test it on new cases. Concerns about using AI for legal judgments are also discussed, such as issues around semantics, bias, and transparency.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd

    Artificial Intelligence in LAW

    Uploaded by

    Madhu
    98 views9 pages
    This document discusses using artificial intelligence and machine learning techniques to help address problems in India's legal system. It outlines issues like a large backlog of pending cases and lack of judicial resources. It then suggests how AI could help with legal tasks like research, contract analysis, and verdict prediction by analyzing past cases and contracts to find patterns and predict outcomes. It proposes an algorithm using long short-term memory to train a model on previous judgments and then validate and test it on new cases. Concerns about using AI for legal judgments are also discussed, such as issues around semantics, bias, and transparency.

    Original Title

    _Artificial Intelligence in LAW

    Copyright

    © © All Rights Reserved

    Available Formats

    PPTX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    This document discusses using artificial intelligence and machine learning techniques to help address problems in India's legal system. It outlines issues like a large backlog of pending cases and lack of judicial resources. It then suggests how AI could help with legal tasks like research, contract analysis, and verdict prediction by analyzing past cases and contracts to find patterns and predict outcomes. It proposes an algorithm using long short-term memory to train a model on previous judgments and then validate and test it on new cases. Concerns about using AI for legal judgments are also discussed, such as issues around semantics, bias, and transparency.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    98 views9 pages

    Artificial Intelligence in LAW

    Uploaded by

    Madhu
    This document discusses using artificial intelligence and machine learning techniques to help address problems in India's legal system. It outlines issues like a large backlog of pending cases and lack of judicial resources. It then suggests how AI could help with legal tasks like research, contract analysis, and verdict prediction by analyzing past cases and contracts to find patterns and predict outcomes. It proposes an algorithm using long short-term memory to train a model on previous judgments and then validate and test it on new cases. Concerns about using AI for legal judgments are also discussed, such as issues around semantics, bias, and transparency.

    Copyright:

    © All Rights Reserved
    Download as PPTX, PDF, TXT or read online from Scribd
    Download as pptx, pdf, or txt
    You are on page 1of 9

    Artificial

    Intelligence in LAW
    Sonakshi Darak, N006 Madhumita Joshi,N015
    Mukesh Patel School Of Technology Management & Mukesh Patel School Of Technology Management &
    Engineering Engineering
    AIML AIML
    sonakshi.darak11@nmims.edu.in madhumita.joshi20@nmims.edu.in
    8080707201 9619014543

    Samiksha Shetty, N038 Saakshi Yadav, N057


    Mukesh Patel School Of Technology Management & Mukesh Patel School Of Technology Management &
    Engineering Engineering
    AIML AIML
    samiksha.shetty70@nmims.edu.in saakshi.yadav 65@nmims.edu.in
    7021454690 9769214424
    PROBLEM STATEMENT & DEFINITION
    ● India now has almost 4 crore pending cases spanning the Supreme Court,
    various high courts and the numerous district and subordinate courts.

    ● It is estimated that India has only 19 judges for every million citizens.

    ● India’s judicial system does have many legal loopholes that hide the corrupt
    nature of the protectors of the law.

    ● Tasks like documentation, analysis etc., used to take a lot of manpower, hence
    resulting in more salary being paid and therefore more financial resources
    being used up.
    PROBLEMS & SUGGESTED
    SOLUTIONS
    Legal Research

    ● Manual Process ● Use NLP to reduce search field by studying


    ● Require Junior associates relevance.
    ● Large set of data ● Use distance measures such as cosine similarity to
    rate relevance of documents.
    ● Cost and time efficient.

    Contract Analysis
    ● Technical Jargon ● Input as above.
    ● Too many clauses. ● Model highlights important parts of contract eg
    ● Requires professional expert to Period for which contract is applied,date for renewal
    understand. etc.
    ● Client remains in dark. ● Highlights context and nature of contract.
    Review of Contracts

    ● Contracts need to be manually reviewed ● Input Lawyer’s & client’s requirements, previous
    and edited,multiple iterations needed contracts to train model.
    ● Human Error ● Input new proposed contract as test case.
    ● Large amounts of data ● Use NLP to either accept/reject rate strength of
    contract or highlight problem areas.

    Verdict Prediction

    ● Large amount of pending cases ● Input previous verdicts of similar


    ● Long drawn cases,legal principles and facts of current
    ● Require lot of resources case.
    ● Subject of matter may be destroyed before ● Use as a precursor step so that all cases
    verdict arrives need not go to trial.
    PREDICTION OF LEGAL
    ●
    OUTCOMES
    This is basically done by building machine learning models to predict
    the outcomes of pending cases, using as inputs the corpus of relevant
    precedent and a case's particular fact pattern.

    ● AI will logically play an important role in assisting litigation funders


    in risk analysis.

    ● The idea is that future legal system could use AI technology to help
    solve disputes without requiring lawyers or the traditional court
    system.
    SUGGESTED ALGORITHM
    1 2 3

    Gather and store data: case number, Preprocess the data and create Apply Long Short-Term
    names of judges, details of all labels. Divide the data in 3 parts Memory(LSTM) on training
    parties, facts about case , final (training, validation and testing) dataset i.e previous cases with
    judgments. known judgement.

    4 5 6

    After model is trained, send model for If it passes with desired accuracy, If validation or testing procedure
    validation. If model passes the then it is implemented. fails then the model is sent again for
    validation and gives right judgement, training.
    then test model with new cases.
    THE DEBATE
    ● Semantics not just Syntax: Consider the context and even
    synonyms
    ● Machine learning models are constrained and task specific.
    ● Bias
    ● Not a transparent method
    ● Outliers

    Will society accept judgement from a machine?

    Can this technology be completely automated?

    Can a Machine replace a Lawyer?


    REFERENCES

    ● Digital Justice and the Use of Algorithms to Predict Litigation Outcomes (Research
    Paper)
    ● AI in legal industry — A case study on predicting judgements through deep learning(
    www.medium.com )
    ● Predictive Analytics For Litigation Case Management (2017) [B y: Jerzy Bala,
    Michael Kellar, Fred Ramberg]
    ● India’s Pending Court Cases On The Rise: In Charts (Bloomberg Quint)
    “
    “Predicting the future isn’t magic,

    It’s Artificial Intelligence ”

    THANK YOU

    You might also like