How AI Is Quietly Changing Income Tax Assessments
There was a time when income tax scrutiny depended largely upon manual selection, human suspicion and occasional luck. Taxpayers often believed that unless their file landed on the “wrong officer’s table”, the chances of detailed scrutiny were relatively low. That era is disappearing rapidly.
Today, Artificial Intelligence (AI), data analytics and automated matching systems are quietly transforming the entire income tax assessment mechanism. The new system does not merely look at the income disclosed in the return. It increasingly compares lifestyle, spending patterns, bank transactions, investments, property purchases, GST data, TDS information, foreign remittances and even digital payment trails.
In short, the Income Tax Department may not know your favourite colour—but it probably knows your spending habits.
The most interesting part is that in many cases, scrutiny selection is no longer driven by “Human Suspicion”. It is increasingly driven by “Data Analytics”. Taxpayers may think this is merely a futuristic concept. However, recent developments suggest otherwise. In one of the most significant technology-driven investigations in recent years, authorities reportedly analyzed nearly 60 terabytes of restaurant billing data covering around ₹2.43 lakh crore of transactions across six financial years. The exercise reportedly indicated possible suppression of sales exceeding ₹70,000 crore through invoice deletions and under-reporting patterns. What is remarkable is that the investigation reportedly began not with a survey or search, but with data analytics. In short, discoveries are increasingly leading to surveys rather than surveys leading to discoveries.
Let us understand how AI is silently changing tax assessments through some practical real-life situations.
1. The “Low Income, Luxury Lifestyle” Problem:
A taxpayer filed ITR showing annual income of ₹6 lakh. However, foreign travel, high-value card spending, luxury purchases and sizeable investments reflected a very different financial profile. AI systems now aggregate and compare these data points automatically. The taxpayer received a notice seeking explanation regarding the source of expenditure and investments. In taxation, Instagram lifestyles and income-tax returns are sooner expected to maintain reasonable coordination.
2. The UPI Trail That Refused to Stay Personal:
A small trader declared turnover of ₹18 lakh under presumptive taxation. However, digital payment data reflected UPI credits substantially exceeding the reported turnover.The taxpayer’s argument was: “Many receipts were personal transfers from friends and relatives.” Unfortunately, AI systems are not very emotional while reading bank statements. Frequent QR-code collections and recurring receipt patterns can now trigger verification. The department increasingly distinguishes between personal transfers and business receipts..
3. The Rent Income Forgotten by the Landlord:
A retired taxpayer honestly disclosed pension income but forgot to include rent received from a tenant who is working in a corporate company. Earlier, such omissions sometimes escaped attention. Today, AIS, TIS and TDS data are automatically matched with the return filed. The taxpayer later received an intimation regarding rent income not reported in ITR. Many taxpayers are surprised to learn that tenants often disclose the landlord’s name, PAN and rent details to their employers while claiming HRA exemption.
4. When the Property Registry Started Talking to the Tax Return:
A salaried employee purchased property worth ₹95 lakh. However, the ITR reflected modest salary income and limited savings history.AI-based systems flagged the transaction because:
• Property registration data,
• Home loan information,
• Stamp duty valuation,
• and Income profile did not appear proportionate.
The department sought explanation regarding the source of funds. Such cases may involve genuine savings or family support. The issue is often documentation rather than tax evasion.
5. The “Invisible” Stock Market Transactions:
AI systems can detect unreported capital gains, derivative losses, dividend mismatches and suspicious loss adjustments. Some taxpayers discover the importance of Schedule CG only after receiving a tax notice.
6. The GST–Income Tax Comparison:
One of the most powerful AI-based assessment tools today is cross-verification between GST returns and Income Tax Returns. Suppose GST turnover is ₹1.8 crore but the turnover reported in the income-tax return is substantially lower or gross receipts in GST do not match banking patterns. Such inconsistencies are increasingly getting flagged automatically. Earlier, different departments behaved like distant cousins.
Today, their databases increasingly behave like close relatives.
7. Online Gaming, Crypto & Digital Income:
Fantasy gaming, crypto trading, influencer income and online earnings have created entirely new tax challenges. Many taxpayers still believe – “Small digital income is invisible.” That assumption is becoming outdated quickly. Platforms increasingly report winnings, withdrawals and transaction data to tax authorities. Some taxpayers celebrate gaming winnings immediately and understand tax implications much later.
8. Fake Deduction Claims Are Becoming Riskier:
One of the biggest behavioural changes created by AI-based scrutiny is in deduction verification. Earlier, inflated deduction claims sometimes escaped verification due to volume limitations. Now, unusual deduction patterns or similar deduction patterns originating from the same IP address can automatically trigger scrutiny indicators.
Examples include:
• Abnormal HRA claims,
• Excessive political donation claims,
• Inflated medical deductions,
• Suspicious charitable contributions,
• and repetitive refund-oriented filing patterns.
A recent refund racket unearthed by the Department provides a perfect example. Data analytics reportedly revealed that a large number of salaried taxpayers claiming similar deductions and refunds were linked to a common e-mail ID and filing pattern. Further investigation allegedly exposed organized filing of returns involving fake and ineligible deduction claims. What earlier may have escaped attention in lakhs of returns was detected because computers are remarkably good at identifying unusual patterns.
Similar analytics have also flagged suspicious political donation claims where deduction patterns and supporting data did not match reported contributions.
9. Refund Processing Is Also Becoming AI-Driven:
AI & Automation work both ways:
• Genuine refunds may come faster,
• But incorrect claims may also get detected faster.
The software remains far more disciplined than taxpayers.
In short, it is speeding up refund processing, mismatch identification and automated communications.
10. The Biggest Change – Assessments Are Becoming “Data Assessments”:
Traditional scrutiny largely depended upon documents physically examined during assessment proceedings. Today, the department may already possess information relating to salary, interest, property transactions, investments, GST turnover, TDS and several other financial activities before the return is even filed. An Income Tax Return is therefore no longer merely a “form”. It is increasingly becoming a “data reconciliation statement”.
Conclusion
Artificial Intelligence is not changing only the speed of tax assessments. It is changing the very philosophy of tax administration. The system is gradually moving from “Detecting Concealment” to “Detecting Inconsistency”. For honest taxpayers, this improves transparency. For careless reporting and shortcuts, the margin of escape is shrinking. After all, in the AI era, every transaction leaves a footprint—and increasingly, those footprints are being followed by machines with memory.
[Views expressed are the personal view of the author. Readers are advised to seek professional advice before taking any decisions. Readers may forward their feedback & queries at nareshjakhotia@gmail.com. Other articles & response to queries are available at www.theTAXtalk.com]
