Source: SCC Online (blog)

In India, Phase III of the E‑Courts Project includes an AI and blockchain component with a separate budget outlay running to 2027, promising intelligent scheduling, case‑prediction, NLP for legal research and document translation, and data security measures, alongside existing tools such as Adalat AI, the Legal Research Analysis Assistant (LegRAA) and SUPACE (the latter currently used only for criminal matters in the Delhi and Bombay High Courts). The piece emphasises that AI systems generate outputs by identifying patterns in large datasets rather than by human value‑based reasoning, and notes specific legal and general AI tools used worldwide and in India, including examples like OLGA, COMPAS, Solution Explorer, Xiao Fa and Victor.

The authors warn that judicial work fundamentally requires the application of human mind to marshal facts and appreciate evidence, and that AI’s apparent efficiency risks substituting independent judicial reasoning with confirmation of machine outputs, creating a “black box” effect and diminishing careful perusal of evidence. They highlight concrete problems such as AI’s tendency to generalise from superficially similar pleadings, the technical limits of OCR and AI on handwritten, poor‑quality, regionally‑language records (for example Pahanis and Adangals), the incomplete adoption of Indian tools, and the proposal that judges should be limited to using explainable AI or otherwise retain strict human control and accountability over judicial decision‑making.