Category: Records, India

“Real-time transcription brings an element of accountability from all the stakeholders. Judges and litigators can no longer make off-the-cuff remarks.

Last week, the Supreme Court witnessed a historic moment in the Indian legal landscape. The proceedings of the case pertaining to the Election Commission’s order on the Shiv Sena party symbol and name were transcribed in real-time using an Artificial Intelligence-based tool. Developed by a Bengaluru-based startup Nomology Technology, the tool attempts to modernise the legal proceedings in the country. 

“Real-time transcription brings an element of accountability from all the stakeholders. Judges and litigators can no longer make off-the-cuff remarks. Now, their actions are under scrutiny. It also makes it easier for litigators to review arguments from the day and avoid repeating themselves, enabling translations of the proceedings, thereby making it more accessible to people who might not speak English.

Most importantly, it makes the proceeding efficient. Currently, each question during a cross-examination might take about five minutes, as the questions and answers from the lawyers and the person testifying is summarised by the judge to the stenographer to be recorded. With real-time transcription, that can be avoided,” says Vikas Mahendra, co-founder of Nomology Technology.

With several years of experience as an arbitration practitioner in England, France and Singapore, Mahendra was taken aback by the lack of proper infrastructure to record court proceedings when he returned home to start a domestic practice in 2015. “A lot of the support mechanisms that were available internationally are not available here. The lack of transcription was glaring. The situation was such that if you wanted transcription, the only choice was to fly down trained transcribers abroad. I sat down with my brother Vinay [Mahendra] and brother-in-law Badarivishal [Kinhal] –both with engineering backgrounds and started discussing what it is that we can do,” he shares. 

Currently, only a minuscule portion of a court proceeding is recorded across the different forums in the country. In lower courts, especially, apart from evidence presentation, virtually nothing else is recorded. Even the miniscule portion that is recorded is often summarised, and as dictated by the judge to the stenographer. Essentially making it hard to piece together the exact happenings of a proceeding.

However, Mahendra feels that recording every single detail from a proceeding is not worthwhile given the sheer volume of cases that pass through the system each year. But, for certain elements, such as witness cross-examination, judgements and evidence presentation, recording and proper transcription can play a crucial role. “We recently had a case where a witness statement was recorded, as summarised by the judge. Later, when the same witness was asked to sign their statement, they refused, saying that’s not what they had said. In such cases, it’s the judge’s word against the witness’, and usually, the judge’s word prevails,” he explains. 

Founded in 2018, Nomology Technology built the transcription engine based on an open-source natural language model. The resulting platform, ‘Technology Enabled Resolution’ has various facets to suit different needs, from live transcription to post-recording transcription, as well as the ability to accurately detect different accents of people across the country and has undergone over 1,500 hours of testing in arbitration hearings. “The platform is constantly evolving. We are also not married on one particular language model,” explains Badarivishal Kinha, operations head and co-founder.

Yet adoption had remained slow until a chance encounter with Chief Justice of India DY Chandrachud in Delhi this month helped change minds. “We decided to transcribe the Delhi arbitration weekend, the flagship event of the Delhi High Court, this month. One of the keynote speakers was Chandrachud, who incidentally, spoke about the need for transcription. That prompted us to speak to him on the sidelines, during which we told him about our work. He asked us to connect with his office. Within a few days, we had set up the infrastructure and did our first transcription,” shares Mahendra, adding they didn’t seek financial compensation for the trial, which helped accelerate the process. 

The Supreme Court recently deployed an Artificial Intelligence -based live transcription tool during court proceedings on an experimental basis. Behind the initiative are city-based startup founders who share how the tech can modernise the legal landscape.

source/content: newindianexpress.com (headline edited)