IIT Kanpur research proposes using AI to predict outcomes of court cases

NEW DELHI: A team of researchers at Indian Institute of Technology (IIT) Kanpur have developed a way of predicting the outcomes of court cases using artificial intelligence (AI). The team from the department of computer science has created a repository of court records, the Indian Legal Documents Corpus (ILDC), which includes 35,000 legal documents from the Supreme Court of India. They intend to use it to train the AI system.

The research paper is titled Indian Legal Documents Corpus based Court Judgement Prediction and Explanation (CJPE). “This is one of the first Indian Legal Corpus of SCI cases created for case prediction and explanation. Moreover, this is an ongoing process and the corpus is being actively built,” said a statement from IIT Kanpur.

The research paper studies the challenges of processing text in legal documents which are usually long and use a specialised legal lexicon. The corpus is annotated with court decisions and a small part of it is annotated with decision explanations. For the prediction of the court judgement, deep learning models have been developed.

Also Read | IIT Ropar launches Data Science and Artificial Intelligence for 12th pass candidates of Punjab

“The corpus is created from publicly available data on the proceedings of the Supreme Court. The AI enabled system only intends to assist legal professionals in their research and decision-making and not replace them. ” the statement further said.

The corpus would also offer a challenging and interesting resource for Legal Natural Language Processing (NLP) researchers. This research was done by including the law professor from West Bengal National University of Juridical Sciences (WBNUJS) as part of the team and it highlights how symbiosis between legal experts and AI researchers could develop practical solutions for the legal system of India.

Write to us at news@careers360.com.