Artificial Intelligence In Decision Making 'Risky', Gujarat High Court's AI-Policy Meant To Preserve Public Trust: CJ Sunita Agarwal

The aim of Gujarat High Court's Policy on use of Artificial Intelligence (AI) in judicial administration is not to prohibit AI but to prevent any situation which may erode public trust or render decision making process vulnerable, said Gujarat High Court Chief Justice Sunita Agarwal on Saturday (May 2). 

Justice Agarwal was speaking at Sikkim High Court's 2-day “National Conclave on Technology and Judicial Education”. Justice Agarwal was chairing a session on the topic "Where AI Can Be Safely Introduced First" along with co-chair Manipur High Court Chief Justice M. Sundar. 

Notably, the Gujarat High Court on April 4 had launched its 'Policy On Use Of Artificial Intelligence In Judicial and Court Administration' which applies to all judicial officers in the state, court staff in high court, legal services authority and district judiciary. The policy prohibits judges from using AI in any form of decision making, adjudication and for preparation of judgments.

Speaking about the High Court's AI policy, Justice Agarwal in her address said:

"Gujarat High Court adopted policy on use of AI in judicial and court administration on April 4. Not with an approach to prohibit or ban AI but as a principal recognition of the need to evolve with changing times. The objective is to prevent any situation which may erode public trust or render decision making process vulnerable. The policy is apportioned against the risk of gradual over reliance on AI and of unintended bias in decision making. While adopting and integrating AI in judicial institutions we understand that this is only a start in a very guarded way balancing skepticism with openness".

Justice Agarwal said that judicial wisdom, according to her, is distinct from intelligence. It is not merely intellectual ability but deliberate application of knowledge of legal principles to impart justice, she added.

"There are risks to judicial independence, and when I am saying so I may put a caveat on it, that I am not criticizing integration of AI. These thought process are for opening discussion to the house. According to me AI comes with many risks when the question is about judicial decision making. A major one is of AI's influence on us due to our complacency. Everyone worries about AI giving wrong answers. But according to me the bigger risk is when AI consistently meets our expectations. We slowly stop questioning and double checking it. Independent judicial thought quietly begins to erode," Justice Agarwal said.

The judge said that a judge's independent freedom means deciding by ones own conscience grounded in legal and constitutional principles which is free from ones own prejudices and biases.

"Allowing the AI to influence that reasoning however certain is against judicial independence itself," the judge added. 

Justice Agarwal further said that as per her understanding AI is more suited for non-adjudicatory functions. She said that AI tools can assist in registry functions involving manual handling such as E filing etc, automated scrutiny and case registration thereby enforcing uniform standards, flagging deficiencies early, timely corrections and drastically cutting filing delays as well as help in identifying case flow bottle necks. 

The judge said that the natural starting point of integration of AI with e-filing and case management software. 

"AI can thus serve as assisted augmentive tool for judiciary. Its integration requires careful oversight and structural deliberations," she said. 

Introducing the Gujarat High Court's AI policy Justice Agarwal also said that a QR code had been created which can be used for downloading the policy. 

The panelists of the sessions were Gauhati High Court judge Justice Michael Zothankhuma, Rajasthan High Court judge Dr. Justice Pushpendra Singh Bhati, and Gujarat High Court judge Justice Justice Nikhil S. Kariel, along with moderator Punjab &Haryana High Court judge Justice Anoop Chitkara.