Manufacturing justice: On the top court and AI use observations

📌 Summary:

• Context: The Supreme Court set aside orders of the NCLT and NCLAT in an insolvency case after finding the NCLT had relied on fictitious AI-generated legal citations — a lapse the appellate tribunal overlooked
• Core argument: AI “hallucinations” in judicial decision-making risk serious miscarriage of justice; the SC likened them to methyl isocyanate (the Bhopal gas tragedy gas) — “invisible, insidious, and catastrophic by the time anyone notices”
• Pattern of interventions: On February 27, the same Bench (Justices P.S. Narasimha and Alok Aradhe) held that a trial court relying on AI-generated fake case laws was not just an error but judicial “misconduct”
• Principle laid down: AI may be an assistive tool to improve efficiency but can never replace independent human reasoning, judicial discretion or professional accountability; a judgment influenced by “an iota” of hallucinated AI material is “no decision in the eyes of law”
• Accountability: Presenting fabricated machine-generated judgments is professional misconduct for advocates and a serious lapse of duty for judges
• Solution / way forward: Draft ‘Regulations for Use of AI in Courts, 2026’ prohibit AI in adjudication, sentencing, deciding bail eligibility and evaluating witness/party credibility; the SC directed the Bar Council of India to set up a committee to frame strict norms and disciplinary action for lawyers citing unverified AI material

🎯 UPSC Relevance: GS2 Polity/Judiciary (and GS3 Technology) — use of AI in the justice delivery system, judicial accountability, and the need for human oversight and regulation of emerging technology.

📝 Prelims Facts:

• Methyl isocyanate (MIC) was the gas involved in the 1984 Bhopal gas tragedy
• Bench: Justices P.S. Narasimha and Alok Aradhe; case involved NCLT and NCLAT (insolvency)
• Draft ‘Regulations for Use of Artificial Intelligence (AI) in Courts, 2026’ open for public consultation; BCI directed to frame norms

🔑 Key Term: AI hallucination — confident but false or fabricated output generated by an AI model (e.g., non-existent legal citations), a core risk when AI is used in judicial reasoning.