The Gujarat High Court has prohibited the use of artificial intelligence in any form of decision-making, legal reasoning, order drafting or judgment preparation, consideration of bail judgments, or substantive court proceedings.
According to the high court’s AI policy announced at a conference of district judicial judges in Gujarat on Saturday, AI should be used to improve the speed and quality of judicial administration rather than as a replacement for judicial reasoning.
The policy states that these technologies “involve significant risks, including hallucinations, bias, breaches of confidentiality, and erosion of judicial independence, and must be managed with care and organizational discipline.”
The report states that by limiting AI to the narrowest possible role, purely anonymized, metadata-driven case assignment and research on legal principles, “human primacy in the delivery of justice is reaffirmed and limited technical assistance is leveraged to reduce administrative imbalances that can delay access to justice.”
“Broadly defined, artificial intelligence shall not be used, directly or indirectly, in any aspect of judicial decision-making, judgment, reasoning, application of law, interpretation of facts, weighing of arguments, determination of rights/liability, sentencing, bail, interim orders, or final judgments,” according to the policy document.
It also states that AI may not be used to discover facts, law, or enforcement orders in any judicial proceeding, nor may AI be used for any task that involves classifying, classifying, organizing evidence, or evaluating or classifying evidence.
According to the policy document, AI cannot be used to create, generate, or substantially constitute a judgment, final order, or binding legal judgment, even if it is later considered by a judge.
It says users cannot enter names, addresses or identifying information of parties, witnesses or advocates, details of pending proceedings or unreported orders, privileged communications or confidential legal strategies, or sensitive personal data.
It states that AI may not be used to generate, fabricate, embellish, or alter evidence in any form; the use of AI-generated citations, case references, or legal provisions without independent verification from an authoritative primary source is prohibited; and the AI is prohibited from drafting, amending, or summarizing office memos or submissions. Judges are personally responsible for all orders, judgments, and findings issued in their name, which cannot be delegated, shared with, or mitigated by the use of AI tools.
Policy emphasizes that all court employees are personally responsible for the accuracy and appropriateness of AI-generated content used in the performance of their official duties. “AI in the judiciary should be designed as a decision support and administrative efficiency tool, not as a replacement for judicial reasoning. With appropriate safeguards such as transparency, human oversight, and protection of sensitive information, AI can significantly enhance case management and improve the speed and quality of justice administration,” the report said.
This policy requires that AI-generated output be reviewed, verified, and held accountable by qualified personnel before it is executed, filed, published, or communicated. AI-generated content, including case law citations and legal references, must be independently verified against trusted primary sources before use.
At the same time, the policy will enable judicial officers and court staff to leverage AI tools to increase productivity, reduce administrative burden, and enhance access to justice, while maintaining judicial independence and the sanctity of judicial decision-making.
The policy allows AI tools to generate code and automate various administrative and productivity tasks, IT department tasks, create presentations and templates for internal training purposes, and draft and improve circulars and notices that are in the public domain.
AI is also permitted to be used for legal research, retrieval and analysis of decisions, identification of precedents, legal interpretation, or any preparatory intellectual work in support of a decision, “provided it is verified using the full human conscience and human mind.”
