Adalat AI, a nonprofit building transcription and document tools for Indian courts, says its software is now used in about 6,000 courtrooms across 11 states, according to reports published online.

The tools aim to reduce the paperwork that clogs India’s court system, with courtroom proceedings transcribed and turned into documents instead of being dictated and typed by hand. The figures come from the organisation itself, and the Semafor summary I could access does not include independent checks of how widely the tools are actually used day to day.

What the reports do not tell us is as important as what they do. Whether judges rely on the transcripts as the official record, how errors in Indian languages and accented English are caught, and what happens to the data are all questions courts and litigants will want answered as these tools spread. But as trust in AI grows, it is sure to make way into the Indian courts. However, there are certain tasks where AI can play a pivotal role in reducing the burden on already paper-mired Indian judicial system. India’s courts have a vast backlog, and any technology that touches the legal record carries more risk than one that only speeds up drafting.

Adalat AI, claimed to be a nonprofit, is building software to cut paperwork in India’s overloaded courts. Its speech-recognition tool is trained on the country’s legal system, and it also offers AI tools to manage cases and court documents. When a witness answers a question, the judge dictates what should be recorded. Both parties check the text on screen, and the transcript becomes the court’s official record.