Artificial Intelligence in Legal Field: Setting Aside of Rs 425 Crore Penalty

Introduction

The Supreme Court has set aside a ₹425 crore penalty imposed by customs after finding that the adjudicating authority relied on fake and AI-hallucinated judicial precedents. This brings us to a fundamental question: with the development of more advanced AI, everybody is getting more involved in using AI for their day-to-day tasks. We know AI has huge potential, and it eases the task, but it should not become a complete substitute for certain fields where human rationality and logic are more important. In this article, we will discuss this incident and argue whether AI should be used in the legal profession or not.

What Actually Happened

In Vijay Ghanshyam Gadiya v. Union of India & Anr., 2026 INSC 947, a Bench of Justices Dipankar Datta and Sheel Nagu dealt with an unusual but increasingly important problem, the use of artificial intelligence in legal adjudication without verification.

The dispute arose from an Order-in-Original passed by the Additional Commissioner of Customs, Surat, imposing a penalty under Section 114 of the Customs Act, 1962 in relation to alleged misdeclaration of natural diamonds as laboratory-grown diamonds. The Supreme Court found that several authorities relied upon in the adjudication order were either non-existent or did not contain the legal propositions attributed to them.

AI Cannot Replace Legal Verification

The judgment is significant because it does not merely concern an erroneous customs order. It addresses the basic responsibility of an adjudicating authority when using AI. Supreme Court has also published a draft regulation regarding the responsible adoption of AI in the judicial system. Additionally, this year the Supreme Court also clarified that any judgment made on AI-generated citations would be void.

Artificial intelligence may assist legal research, but the authority deciding a case remains responsible for verifying the existence and correctness of every precedent relied upon. The Court consequently treated reliance upon fabricated legal material as a serious defect in the decision-making process.

Why the Judgment Matters

The decision is particularly relevant for courts, tribunals, government departments and lawyers increasingly using generative AI for legal research. The Supreme Court has effectively drawn a line between using AI as an assistive research tool and treating AI-generated material as an authority without verification. The legal proposition is therefore wider than customs law. A judicial or quasi-judicial decision cannot be sustained merely because its reasoning appears legally structured if the authorities supporting that reasoning do not actually exist.

Should AI Be Used in the Legal Profession

Yes, AI should be used in the legal profession, but with the caution that the Hon’ble Supreme Court stated that humans must verify it. As we know, AI is trained on information available in the public domain or some specific information by creator; it is not truly unbiased because the raw data itself contains certain biases which is propagated through AI. I have previously discussed in my research paper, the impact of AI on freedom of speech and expression, where I pointed out how AI can spread propaganda and manipulate information.

Conclusion

The judgment sends a straightforward message: technology can assist legal reasoning, but it cannot replace legal verification. Advocates, judges and other tribunal and judicial bodies should verify through legal research that the judgment exists and that it conveys the same rationale as it is cited for. Otherwise, we will fall into the laps of giant corporation whimsical analogies.

FAQs

• Can AI-generated case laws be relied upon in legal proceedings?
• What happens when an adjudicating authority relies on a non-existent judgment?
• What does Section 114 of the Customs Act provide?
• Can an entire administrative order be invalidated because of false legal precedents?