Source: Media nama
The Gujarat High Court has quashed a GST case after an officer used AI-generated case law that did not exist. Justices A.S. Supehia and Vaibhavi D. Nanavati delivered the judgment on August 20, 2026.
The petitioner, Faiz Enterprise, challenged a show-cause notice, GST cancellation order, revocation rejection, and appellate order. The dispute arose after the officer relied on judgments that were non-existent, wrongly cited or irrelevant. The Court first flagged the issue on August 13. It then examined the officer’s apology and departmental response before quashing the proceedings.
What the judgment said: The court said the officer had relied exclusively on AI-generated case law. It noted that the cited judgments were non-existent or irrelevant. “It appears that the impugned order has been passed by the State Tax Officer by placing reliance exclusively on the AI-generated case laws.”
The officer, Devang Arvindkumar Yadav, tendered an unconditional apology. He said his lack of experience caused the mistake. He admitted to using AI to draft the order and cite the judgments. He also said he attended an AI training programme on August 18. The Court directed strict compliance with the departmental AI instructions. It said, “any violation of the instructions would amount to contempt of this Court”.
The Court quashed the show cause notice, cancellation order, revocation rejection and appellate order. It directed the department to issue a fresh notice. It ordered the authority to consider the petitioner’s defence and pass a reasoned order. The Gujarat State Tax Department issued its AI instructions on August 18, after the Court flagged the use of non-existent AI-generated case law.
What the State Tax Department’s AI rules say: The Court ordered the department to follow the August 18 instructions scrupulously. It said any violation would amount to contempt. The rules require:
- Independent verification: Officers must verify AI-identified laws and cases through primary official sources.
- Citation checks: Officers must verify citations, relevance, authenticity, and current legal status.
- Original text: Officers must locate original paragraphs when AI paraphrases legal reasoning. They must reproduce them verbatim.
- Taxpayer submissions: Officers must consider judgments cited by taxpayers. They must record reasons for rejecting their relevance.
- Human oversight: Officers must maintain human oversight when using AI or other technological tools.
- Independent reasoning: AI can assist reasoning but cannot replace the officer’s application of law and facts.
Why does this matter? The Gujarat case shows why AI accountability needs to be embedded in government processes. Government officials make decisions that can alter businesses’ rights, liabilities and finances. Here, a State Tax Officer used AI-generated, non-existent judgments while drafting a GST order. The HC then quashed the proceedings and ordered fresh adjudication.
The problem is broader. In July, the Supreme Court set aside NCLT and NCLAT orders after tribunals relied on six defective AI citations. In the same month, the Delhi HC also set aside a tax appellate order containing six non-existent judgments. The officer had inserted citations supplied by an inspector without verification. Meanwhile, a Jammu and Kashmir trial court judge admitted to using AI without checking citations.
Share this content: