SC declines PIL on regulating AI use in judiciary, allows petitioner to submit suggestions

The Supreme Court on Friday declined to entertain a PIL seeking regulation of the “unregulated” use of artificial intelligence (AI) and machine learning (ML) tools in courts, observing that the concerns raised were valid but could be addressed more effectively on the administrative side rather than through judicial orders.

A bench of Chief Justice Surya Kant and Justice Joymalya Bagchi heard senior advocate Anupam Lal Das , appearing for petitioner Kartikeya Rawal , who approached the court citing the growing risks posed by AI-generated content in judicial work. The petitioner argued that AI tools were increasingly producing non-existent case laws and fabricated judicial precedents , some of which were being unknowingly cited by lawyers and even reflected in lower court orders . He said this trend threatened the integrity of judicial processes and warranted regulatory safeguards from the apex court.

Das warned that AI “hallucinations” false or invented judgments could mislead courts if not regulated, noting that certain trial courts had already cited fake Supreme Court rulings generated by AI tools. He pointed to initiatives such as the Kerala High Court’s structured mechanism on AI use and the Supreme Court’s white paper on AI as examples of ongoing efforts, arguing that a uniform framework was needed nationwide.

While acknowledging these risks, Chief Justice Kant said the judiciary was conscious of the pitfalls of unchecked AI use and had already begun addressing them internally. He emphasised that both judges and lawyers have future responsibilities when using AI tools, including mandatory verification of all AI-generated citations and ensuring that no unverified content influences judicial reasoning. “Judges must cross-check. This is part of judicial academy training and is being taken care of. With time, the Bar will learn and we will also learn,” he said.

The CJI reiterated that AI may assist with administrative tasks but cannot replace or overpower judicial decision-making , stressing the need for cautious and responsible use. He said judicial academies and Bar bodies would continue to train officers and advocates on the ethical use of AI, including recognising fabricated content and maintaining human-centric adjudication.

Stating that the matter was better suited for administrative action rather than judicial directions, the bench invited the petitioner to submit recommendations directly to the Supreme Court. “Someone with sincere intentions is most welcome to give us suggestions. You can mail them to us,” the CJI said.

Sensing the inclination of the bench, Das sought permission to withdraw the PIL, which was granted.