Supreme Court warns AI-generated judgments will amount to misconduct

NEW DELHI: The Supreme Court on Monday warned that judicial orders based on AI-generated, non-existent judgments would amount to misconduct and not just decision-making errors, signaling serious concerns over the use of artificial intelligence in court proceedings.A bench of Justices PS Narasimha and Alok Aradhe said the issue will be examined in detail and notices will be issued to Attorney General R Venkataramani, Attorney General Tushar Mehta and the Bar Council of India, PTI reported. Senior advocate Shyam Divan has been appointed to assist the court.“We recognize that the trial court deployed an artificial intelligence-generated purported verdict that does not exist, is false, or is synthetic, and seek to review its consequences and liability because of its direct impact on the integrity of the trial process,” the judge said in the Feb. 27 order.According to PTI, “First of all, we must state that a decision based on such a non-existent and spurious so-called verdict is not an error in decision-making. It will be a form of misconduct and will have legal consequences. We must examine this issue in more detail, which is compelling.”The issue came up during a plea hearing challenging a January order of the Andhra Pradesh High Court in a suit seeking an injunction. The Supreme Court noted that the trial court relied on certain judgments in rejecting objections to the Advocate Commissioner’s report. The petitioners argued that the judgment cited did not exist and was generated by artificial intelligence.The High Court acknowledged that the judgment cited was AI-generated and recorded a caveat, but continued to rule on the merits and dismissed the civil modification application. The petitioners then moved the Supreme Court.The Supreme Court issued a notification directing that the trial court “shall not proceed with the hearing on the basis of the Advocate Commissioner’s report” until the special leave application is disposed of, and published the matter on March 10. In another hearing on February 17, a bench headed by Chief Justice Surya Kant, while hearing a PIL seeking guidelines on political speeches, also expressed concern over petitions filed by lawyers drafted using AI tools citing non-existent cases like “Mercy and Humanity”.