The Core of the Proposal
A recent report from the Parliamentary Standing Committee on Finance has put forward a significant recommendation to tackle the growing issue of delayed income tax refunds. The panel, chaired by BJP leader Bhartruhari Mahtab, suggested the implementation
of an AI-based system to risk-score tax filings. The core idea is to automatically categorise returns into low, medium, and high-risk profiles. This comes in response to a sharp increase in the number of refunds taking more than 90 days to process, which climbed to nearly 27 lakh cases in the 2025-26 financial year. The committee believes this technological intervention can make tax administration more targeted and less intrusive for the majority of honest taxpayers.
How Would AI Risk-Scoring Work?
The proposed system would function by having an AI algorithm analyse vast amounts of data to assess the probability of a fraudulent or inaccurate claim. This data would include not just the information in your current tax return, but also your past filing history, data from your Annual Information Statement (AIS), GST returns, bank transactions, and property purchases. Based on this 360-degree profile, the AI would assign a risk score. A low-risk score would signal a compliant taxpayer, whose refund could potentially be processed almost instantly. Conversely, a high-risk score would flag the return for detailed manual scrutiny by a tax officer, leading to delays.
The Promise: Faster Refunds and Better Efficiency
The primary goal behind this proposal is to benefit the vast majority of compliant taxpayers. By automatically identifying and fast-tracking low-risk returns, the system aims to drastically cut down waiting times for refunds. This would alleviate the cash-flow stress that delays can cause for salaried individuals and small businesses. For the Income Tax Department, it means human resources can be channelled away from reviewing every single file and instead focused on a smaller number of high-risk cases where the chances of tax evasion or significant errors are much higher. This is a move towards making enforcement more precise and less burdensome for the general public.
The Concerns: Algorithmic Bias and Transparency
While the promise of efficiency is attractive, the proposal also brings significant concerns. Experts worry about the 'black box' nature of AI, where the logic behind a decision might not be clear. What if an algorithm develops biases against certain professions, income brackets, or geographical regions, unfairly flagging them as high-risk? This could lead to honest taxpayers getting stuck in a cycle of delays and scrutiny without a clear explanation. There are also calls for strong governance guardrails, including the need for an AI ombudsperson to review contested decisions and ensure fairness in the system. Without transparency and accountability, an AI-driven system could erode taxpayer trust.
An Evolution, Not a Revolution
It's important to note that the Income Tax Department is already using AI and data analytics quite extensively. 'Project Insight', which became fully operational in 2019, uses data to create 360-degree profiles of taxpayers to detect inconsistencies and encourage voluntary compliance. The department also uses a 'nudge' strategy, sending automated emails and SMS alerts to taxpayers based on AI analysis. According to the Central Board of Direct Taxes (CBDT), these AI tools have already helped recover significant tax revenue. Therefore, the parliamentary panel's proposal is not about introducing AI from scratch, but rather about formalising and expanding its role specifically for the refund process.
