Use Of AI-Generated Fake Citations Doesn't Vitiate Order If Correct Law Applied: Andhra Pradesh High Court
The Andhra Pradesh High Court has observed that mere mention of non-existent citations generated by Artificial Intelligence (AI) would not vitiate an order, if the law considered in the order is correct and there is no fault in the application of the law to the facts of the case.The Court observed so while dismissing a challenge to an order of the Trial Court, where the judicial officer...
The Andhra Pradesh High Court has observed that mere mention of non-existent citations generated by Artificial Intelligence (AI) would not vitiate an order, if the law considered in the order is correct and there is no fault in the application of the law to the facts of the case.
The Court observed so while dismissing a challenge to an order of the Trial Court, where the judicial officer had, while upholding an advocate commissioner's report, relied on citations generated by AI without subsequent verification, resulting in the inclusion of non-existent authorities in the order.
The judicial officer had nonetheless accorded reason while upholding the report, noting that the report is a valuable piece of evidence, which can be scrutinised in light of other pieces of evidence and shall be considered at the trial stage, unless shown to be vitiated by bias or misconduct.
The said order was challenged before the High Court on the ground that citations relied on by the Trial Court did not exist and thus the order was liable to be set aside.
Dismissing the challenge, Justice Ravi Nath Tilhari held that while the citations may be non-exist, if the Trial Court has considered correct principles of law, and its application to the facts of the case is also correct, then “mere mentioning of incorrect or non-existent rulings/citations in the order cannot be a ground to set aside the order. If the principle of law applied is not the law of the land or its application in a given case is faulted because of relying on non-existent rulings generated by AI, then the case for interference would be made out.”
Noting that the judicial officer gave a reasoned order supported by law, the Single Judge concluded,
“The reasons recorded in the order and the view taken by the learned Trial Court, is perfectly justified. It has the support of law. The commissioner's report is a piece of the evidence. It is to be considered at the final hearing subject to the objections raised and the evidence on record. The petitioners may raise objections to the report, for which, the learned Trial Court has also observed viz., by way of cross examination. The order therefore, does not cause any prejudice to the petitioners at this stage. The opportunity has been kept open to the petitioners to raise the correctness of the report on evidence. This Court does not find any illegality in the impugned order. Merely because non-existing citations were referred due to the use of Artificial Intelligence Tool, without due verification, the order would not be vitiated, when the law stated in the order and its application is correct…”
However, the Court cautioned against the unregulated use of AI tools in legal research and judicial writing, holding that though AI tools are meant to increase efficiency, their unregulated use in judicial parlance give rise to serious concerns– including violations of privacy and damage to the confidence and trust in judicial decision-making.
Reprimanding the practice of blind reliance on AI tools without meaningful human oversight to verify the accuracy of assertions they generate, the Single Judge explained,
“The people who employ AI for legal research should rigorously scrutinize its outputs, including the authorities cited. Such tools may lack access to the complete body of relevant law, may not fully comprehend the query posed, or may overlook material authorities. In such circumstances, AI systems can produce responses that appear persuasive yet are factually or legally incorrect. More concerningly, they may fabricate authorities or cite existing cases that are irrelevant to the issue under consideration.”
Justice Tilhari further observed that application of actual intelligence should be preferred over AI, and AI tools should be used with caution and wisdom. Trial Courts were also advised to remain vigilant and to act with judicial application of mind while using AI in the process of judicial decision-making in order to ensure that judgments are premised on correct legal principles.
“The submission of fictitious opinions gives rise to multiple harms: the opposing party is compelled to expend time and resources in exposing the falsity; valuable judicial time is diverted from legitimate adjudicatory functions; and litigants may be deprived of arguments grounded in genuine judicial precedents. There is also the serious risk of damage to the reputation of the judges and the courts whose names are falsely attributed to fictitious opinions, as well as to parties associated with fictional conduct. Such practices foster mistrust toward the legal profession and the Justice delivery system.”, the Court added.
Accordingly, the Court dismissed the petition.
Case Details:
Case Number: CIVIL REVISION PETITION NO:2487 OF 2025
Case Title: GUMMADI USHA RANI v. SURE MALLIKARJUNA RAO