AI-Powered Tax Governance in India

India continues to face a persistent fiscal challenge of a low tax-to-GDP ratio and widespread tax evasion. Between 2001 and 2022, the average tax-GDP ratio remained around 16.36%, which is relatively low compared to many emerging economies. In addition, it is estimated that nearly 4.3% of tax revenue is lost annually due to evasion.

To address these challenges, India has increasingly adopted Artificial Intelligence (AI) in governance, particularly in tax administration. A key initiative in this direction is the Income Tax Department’s Project Insight, which aims to improve revenue mobilisation, strengthen voluntary compliance, and ensure fairness in enforcement.

Architecture of Project Insight

Project Insight was launched in 2017 and became fully operational by 2019. It is designed as a data-driven and AI-enabled tax compliance system to identify high-risk tax evasion cases and improve overall compliance.

Income Tax Transaction Analysis Centre (INTRAC)

The core component of Project Insight is the Income Tax Transaction Analysis Centre (INTRAC). It uses advanced data analytics and AI tools to integrate and analyse information from multiple sources, including:

• Banking transactions

• Financial institutions

• Property records

• GST filings

• High-value financial transactions

This integration helps create a 360-degree profile of taxpayers, allowing authorities to identify mismatches between declared income and actual financial behaviour.

This enables data-driven detection of tax evasion patterns that were previously difficult to trace manually.

Compliance Management System and NUDGE Strategy

Project Insight is supported by a Compliance Management Centralised Processing Centre, which promotes voluntary compliance using behavioural insights.

It follows the NUDGE strategy (Non-intrusive Usage of Data to Guide and Enable), under which:

• Taxpayers receive SMS or email alerts

• They are encouraged to voluntarily correct discrepancies in returns

• They can file revised returns without coercive action

This approach shifts tax enforcement from punitive action to behavioural compliance and voluntary correction.

Benefits of AI-Driven Tax Systems

AI-based tax governance provides several important advantages:

1. Improved Risk Profiling

AI helps in identifying high-risk taxpayers and suspicious transactions more accurately.

2. Better Prioritisation of Cases

Authorities can focus on cases involving large-scale evasion or complex financial irregularities.

3. Automation of Processes

Routine tasks such as data matching and verification are automated, improving efficiency and reducing manual workload.

4. Improved Taxpayer Services

AI enables chatbots, digital assistance, and faster query resolution, improving taxpayer experience.

5. Fraud Detection

AI systems help detect complex and hidden patterns of tax evasion, strengthening enforcement.

Measurable Outcomes and Global Context

The implementation of India’s AI-based tax system through Project Insight has produced significant measurable outcomes in improving tax compliance and revenue collection.

Since 2020–21, more than one crore revised income tax returns have been filed, resulting in an additional revenue collection of around ₹11,000 crore. This shows that AI-enabled monitoring has helped expand voluntary compliance.

Targeted NUDGE-based campaigns have also improved disclosure quality. These interventions led to better reporting of foreign income and overseas assets, while corrections related to false deductions resulted in ₹963 crore worth of corrections and ₹410 crore in additional tax payments.

Administrative efficiency has also improved significantly, as the refund processing time has reduced from 93 days to 17 days, making the system faster and more taxpayer-friendly.

AI tools have further enabled the detection of large-scale tax evasion, including identification of nearly ₹70,000 crore in suppressed restaurant sales, achieved through methods such as invoice deletion and data manipulation.

These outcomes demonstrate that AI and big data analytics are highly effective in detecting complex and hidden forms of tax evasion.

Global Context

At the global level, countries such as Australia, Italy, the United Kingdom, and the United States have adopted similar AI-based tax governance systems. This reflects the growing global reliance on data-driven tax administration.

India’s adoption of Project Insight aligns with these international practices and positions it as a leader in digital public finance and AI-enabled governance systems.

Concerns and Structural Risks

Despite its benefits, AI-powered tax governance presents several important challenges and risks.

1. Data Quality Issues

AI systems depend heavily on accurate and complete datasets. However, complex financial arrangements such as irregular income patterns or joint family finances may lead to false positives, where genuine taxpayers are wrongly flagged.

Poor data quality can lead to misclassification and unnecessary compliance burden on taxpayers.

2. Algorithmic Bias

AI models trained on historical datasets may unintentionally reinforce existing socio-economic inequalities, leading to biased targeting of certain groups.

A global example of this risk is the Dutch childcare benefits scandal, where algorithmic bias led to wrongful penalties and large-scale injustice.

3. Lack of Transparency and Due Process

A major concern is the lack of explainability in AI decisions. Taxpayers often do not understand:

• Why they were flagged

• How decisions were made

• What data influenced the decision

The absence of a strong human-in-the-loop system and transparent appeal mechanism weakens procedural fairness and accountability.

4. Data Privacy and Cybersecurity Risks

The integration of large volumes of sensitive financial data increases risks such as:

• Data breaches

• Cyberattacks

• Misuse of confidential financial information

  Strong data protection and cybersecurity frameworks are essential to ensure trust in AI-driven systems.

5. Institutional Gaps

India currently lacks:

• An AI ombudsperson for grievance redressal

• Mandatory algorithmic audits

• Systems to track false positives and appeal outcomes

Conclusion

The AI-driven tax governance system under Project Insight demonstrates the transformative potential of technology in improving tax compliance, efficiency, and revenue mobilisation.

However, its long-term success depends on maintaining a careful balance between technological efficiency and principles of fairness, transparency, and accountability.

Strengthening data governance, privacy protection, algorithmic audits, and independent oversight mechanisms is essential to address existing risks.