AI should serve justice, not rule it: Justice Manmohan
While delivering former law Minister Dr HR Bhardwaj Memorial lecture on the topic “The emergence of AI and its implications for Justice delivery and legal systems” on Sunday, Justice Manmohan underscored the Supreme Court’s proactive measures, including the deployment of tools like SUVAS for vernacular translations and AI-based speech-to-text systems for transcribing oral arguments. He also highlighted “LegRAA,” a generative AI tool currently in its pilot phase, trained on Indian case law. Furthermore, he noted the Chief Justice of India’s recent unveiling of the “one case one data” initiative and the AI chatbot “Su Sahay”. He reiterated the institution’s core governing mantra: “Assistive, not determinative”.
Highlighting the rapidly evolving integration of Artificial Intelligence in the legal landscape, Justice Manmohan emphasised that technology must remain an assistive tool rather than a substitute for human judgment. Delivering a lecture on AI, he stated that while technological advancements reshape justice administration, they demand a rigorous reckoning regarding fairness and reliability.
Cautioning against blindly copying global models, Justice Manmohan contrasted the efficiency of China’s “Internet Courts” with their lack of human empathy. He also cited the American COMPAS tool as a cautionary tale of “black box” algorithms digitising systemic social biases. Turning to ethical hazards, he warned of “hallucinations”, citing a New York case where a lawyer submitted completely fabricated, AI-generated precedents. This, alongside the Indian Supreme Court’s recent observations in Usha Rani’s case, shifts the traditional legal “Duty of Candour” towards a stringent “Duty of Verification”.
The lecture further addressed data privacy and algorithmic biases. Justice Manmohan praised sectoral regulators like the RBI and SEBI for moving the needle from individual consent to institutional safety. He noted that the RBI’s “Seven Sutras” mandate human-overridable decisions to counteract hidden AI credit-scoring inferences, effectively re-humanising contracts.
Advocating for open-source and transparent algorithms within the Indian judiciary, he warned against a digital divide that could create a two-tier justice system favouring the urban elite. Justice Manmohan concluded by calling for robust regulatory frameworks and university courses on AI ethics. Urging the legal fraternity to act as wise architects, he remarked, “Build the digital seatbelts before the accidents happen”.
