AI Tools Uncover ₹68 Lakh Capital Gains Tax Fraud in Hyderabad

Current position — reviewed 22 August 2026

The case below is reported as it stood when the department acted on it. The capital gains law has since changed, so read the numbers in that light.

• Indexation benefit was withdrawn for transfers made on or after 23 July 2024; long-term capital gains on immovable property are now computed without indexation at the current rate, subject to the transitional relief available to individuals and HUFs for property acquired before that date.
• Any cost of improvement claimed must be supported by dated invoices, bank payment trails and, where relevant, municipal approvals — unsupported claims are exactly what the department's analytics now isolate.
• From 1 April 2026 assessments run under the Income-tax Act, 2025; the substantive computation and evidence expectations described here continue to apply.

Hyderabad, July 2025 – In a striking example of how technology is reshaping tax enforcement, the Income Tax Department has unearthed a ₹68 lakh capital gains tax fraud in Hyderabad using artificial intelligence (AI) and digital forensic tools.

The case not only highlights the growing sophistication of fraud attempts but also underlines how AI-based systems are becoming a powerful ally for tax authorities.

The Case: Property Sale and Suspicious Claims

The taxpayer at the center of the case had sold a property for ₹1.4 crore. Normally, capital gains tax is calculated by deducting the indexed cost of acquisition and any genuine cost of improvement from the sale price.

However, in this case, the individual significantly reduced taxable gains by claiming:

• ₹73 lakh as indexed purchase cost
• ₹68.7 lakh as “cost of improvement”

As a result, the declared long-term capital gains (LTCG) came down to just ₹24,774 – an abnormally low figure given the scale of the transaction.

To support the claim, the taxpayer submitted several photocopied expenditure bills supposedly dated between 2002 and 2008. One particular bill, dated July 6, 2002, raised suspicion.

How AI Flagged the Fraud

The Income Tax Department employed AI-powered forensic analysis to examine the documents. Unlike manual inspection, AI tools can analyze fonts, file metadata, ink patterns, and inconsistencies that human eyes often miss.

In this case, the AI flagged the July 2002 bill for one simple but critical reason: it was typed in the Calibri font.

• The Calibri font was designed between 2002 and 2004.
• It only became available for public use in 2006.
• It was set as the default font in Microsoft Office in 2007.

Since the bill was supposedly created in 2002, its use of Calibri clearly proved that the document was fabricated.

This revelation exposed the ₹68 lakh false improvement cost, which had been artificially inflated to reduce taxable gains.

Taxpayer’s Response

When questioned, the taxpayer claimed that the bills were discovered in an old folder belonging to a deceased family member. The individual expressed uncertainty about the authenticity of the documents but initially used them to file returns.

After the fraud was detected, the taxpayer was forced to withdraw the false claim and file a revised return, paying taxes on the actual capital gains.

The Role of AI in Tax Enforcement

This case highlights how AI is increasingly becoming a game-changer in tax enforcement. Traditionally, fraudulent documents relied on the assumption that manual verification would miss small details. With AI, even subtle inconsistencies—like a font released years later or a watermark mismatch—can expose manipulation.

Globally, tax authorities are investing in:

• AI-based document forensics to spot fake bills, invoices, and contracts.
• Data analytics to identify abnormal patterns in income declarations.
• Machine learning algorithms to detect fraud networks and shell company linkages.

India’s tax department is now actively deploying these tools to strengthen compliance and reduce tax evasion.

Lessons for Taxpayers

This incident carries important lessons for taxpayers:

1. Fabricated documents will not stand scrutiny – Modern forensic tools can detect fraud within seconds.
2. Technology ensures fairness – Honest taxpayers need not fear, but fraudulent claims are increasingly easy to detect.
3. Transparency is essential – Maintaining genuine purchase and improvement records is the safest way to comply with tax laws.

Conclusion

The Hyderabad case is a strong reminder that in the age of artificial intelligence, dishonesty in tax filing has little room to hide. A seemingly minor detail—a font not available in 2002—was enough to unravel a multi-lakh fraud attempt.

By leveraging AI and digital forensics, the Income Tax Department is not only protecting government revenue but also building a more transparent and equitable tax system.