The Gujarat Excessive Court docket has prohibited using Synthetic Intelligence for any type of decision-making, judicial reasoning, order drafting or judgment preparation, bail sentencing issues, or any substantive adjudicatory course of.
In response to the excessive courtroom’s AI coverage, unveiled on Saturday (April 4, 2026) at a convention of district judiciary judges in Gujarat, AI ought to be used to enhance the pace and high quality of justice supply, slightly than as a substitute for judicial reasoning.
As per the coverage, these applied sciences “carry substantial dangers — together with hallucinations, bias, confidentiality breaches, and erosion of judicial independence — that have to be managed with care and institutional self-discipline”.
It says that by confining AI to the narrowest conceivable position, purely anonymised, metadata-driven case allocation and analysis of authorized rules, “human supremacy in justice supply is reaffirmed and restricted technological help is harnessed to cut back administrative imbalances that may in any other case delay entry to justice”.
“Broadly, Synthetic Intelligence shall not be used — straight or not directly — for any side of judicial determination, adjudication, reasoning, utility of regulation, interpretation of info, weighing of arguments, willpower of rights/liabilities, sentencing, bail, interim orders, or remaining judgment,” as per the coverage doc.
Additionally, AI shall not be used to seek out info, regulation, or operative order in any judicial continuing, nor shall or not it’s used for sorting, classification of proof, group of evidentiary materials, or any activity involving analysis or categorization of proof, it says.
AI can’t be used to creator, generate, or considerably compose any judgment, remaining order, or binding authorized ruling, even when subsequently reviewed by a decide, based on the coverage doc.
One can not enter names, addresses, or figuring out info of events, witnesses, or advocates, particulars of pending proceedings or unreported orders, privileged communications or confidential authorized methods, and delicate private information, it says.
AI shall not be used to generate, fabricate, embellish, or alter proof in any type, utilizing AI-generated citations, case references, or statutory provisions with out unbiased verification from authoritative main sources can also be prohibited, as is drafting, correcting or summarizing any workplace word or submission, it says. A decide is personally answerable for each order, judgment, and commentary issued beneath their identify, and this can’t be delegated, shared with, or diminished by way of any AI instrument, it says.
Each courtroom officer is personally answerable for the accuracy and appropriateness of any AI-generated content material used within the efficiency of their official duties, stresses the coverage “AI in judiciary ought to be designed as a decision-support and administrative effectivity instrument, not as a substitute for judicial reasoning. With correct safeguards — corresponding to transparency, human supervision, and safety of confidential info — AI can considerably strengthen case administration and enhance the pace and high quality of justice supply,” it states.
The coverage requires a certified human officer to evaluation, confirm, and be answerable for any AI-generated output earlier than it’s acted upon, filed, revealed, or communicated. AI-generated content material, together with case citations and statutory references, have to be independently verified towards authoritative main sources earlier than use, it says.
On the similar time, the coverage permits judicial officers and courtroom workers to leverage AI instruments to enhance productiveness, cut back administrative burden, and improve entry to justice, whereas preserving judicial independence and the sanctity of judicial decision-making.
The coverage permits AI instruments for numerous administrative and productiveness duties, code era or automation for IT division duties, creating displays or templates for inner coaching functions, and drafting and bettering circulars and notices, details about which is within the public area.
AI can also be permitted for use for authorized analysis, retrieval or evaluation of judgments, identification of precedents, statutory interpretation, or any preparatory mental work supporting adjudication, “however with all human conscience and topic to verification by making use of thoughts”.
Revealed – April 05, 2026 07:54 am IST