Context

• A persistent challenge in India’s fiscal system is its low tax-GDP ratio and widespread tax evasion.
• Between 2001 and 2022, the ratio averaged 16.36%, among the lowest for emerging economies, while about 4.3% of revenue is lost annually due to evasion.
• With the rise of Artificial Intelligence (AI) in governance, India has begun leveraging technology to improve tax administration.
• The Income Tax Department’s Project Insight (PI) represents a major step toward strengthening revenue mobilisation, enhancing voluntary compliance, and ensuring fairness in enforcement.

The Architecture and Benefits of Project Insight

• Launched in 2017 and fully operational by 2019, Project Insight is designed to improve tax compliance and reduce high-risk evasion.
• Its core component, the Income Tax Transaction Analysis Centre (INTRAC), uses data analytics and AI to process information from banking systems, financial institutions, property records, GST filings, and high-value transactions.
• This creates a 360-degree taxpayer profile, enabling authorities to detect mismatches between declared income and actual financial behaviour.
• The Compliance Management Centralized Processing Centre complements this system by encouraging behavioural compliance through the NUDGE strategy (Non-intrusive Usage of Data to Guide and Enable).
• Taxpayers receive reminders via SMS or email to correct discrepancies, allowing them to file revised returns voluntarily. This reduces reliance on coercive enforcement.
• AI-driven tax systems offer multiple benefits. They enable accurate risk profiling, help prioritise cases based on scale and complexity, and automate routine processes, increasing efficiency.
• They also improve taxpayer services through chatbots, timely assistance, and fraud prevention mechanisms. Overall, AI enhances both enforcement capability and user experience.

Measurable Outcomes and Global Context

• Project Insight has delivered significant outcomes. Since 2020–21, over one crore revised returns have generated an additional ₹11,000 crore.
• Targeted NUDGE campaigns have led to substantial corrections in foreign income and overseas assets disclosures, with high compliance rates. Similarly, campaigns addressing false tax deductions resulted in corrections worth ₹963 crore and additional tax payments of ₹410 crore.
• Administrative efficiency has improved, with refund processing time reduced from 93 to 17 days.
• AI tools have also uncovered large-scale evasion, including ₹70,000 crore in suppressed restaurant sales through methods such as invoice deletion and data manipulation.
• These outcomes demonstrate the effectiveness of big data and AI in identifying sophisticated fraud.
• Globally, countries like Australia, Italy, the United Kingdom, and the United States have adopted similar systems, confirming the relevance of AI-driven tax governance.
• India’s approach aligns with these international practices, positioning it as a leader in digital public administration.

Concerns and Structural Risks

• The first concern is data quality; AI systems depend heavily on accurate data, but complex financial situations, such as irregular income or joint family finances, can generate false positives, placing the burden on taxpayers to justify legitimate transactions.
• Second, algorithmic bias is a significant issue. AI models trained on historical data may reinforce existing socio-economic disparities, leading to unequal targeting.
  • Experiences like the Dutch childcare benefits scandal illustrate the dangers of biased systems.
• Third, the lack of explainability and due process raises questions of fairness.
  • Taxpayers must understand why they are flagged, how decisions are made, and have access to a transparent appeal mechanism.
  • The absence of a clear human-in-the-loop system for critical decisions can undermine trust.
• Fourth, concerns about data privacy and cybersecurity are critical. The aggregation of sensitive financial data increases vulnerability to breaches and misuse.
• Finally, institutional gaps persist; India lacks an AI ombudsperson, formal algorithmic audits, and systems for reporting false-positive rates or appeal outcomes. Without these safeguards, accountability remains limited.

Conclusion

• Project Insight demonstrates the transformative potential of AI in improving tax governance, enhancing compliance, and boosting revenue.
• However, its long-term success depends on balancing technological efficiency with transparency, accountability, and fairness.
• Robust safeguards, including independent oversight, better data governance, and clear legal frameworks, are essential.
• Strengthening trust between taxpayers and authorities is critical to sustaining compliance.
• India must ensure that its transition to AI-driven systems promotes not just efficiency, but also equity and justice, creating a model of ethical AI governance in public finance.