Insights

New Federal Tax Law: R&D Deduction Changes

Richard Levychin, CPA CGMA Featured on Podcast “Startups With Stu” Discussing Exiting Your Company

Potential Opportunities And Benefits To Taxpayers From Recent IRS Cutbacks

The GENIUS Act: A New Era for Stablecoin Regulation in the U.S. Introduction

How the IRS is Leveraging Artificial Intelligence
to Transform Tax Administration

The Internal Revenue Service (IRS) is increasingly harnessing the power of artificial intelligence (AI) to enhance tax administration, streamline operations, and combat tax evasion. With a growing tax gap—estimated at $688 billion annually for tax year 2021—the IRS is deploying AI to improve compliance, optimize resources, and deliver better taxpayer services. This article explores the key ways AI is transforming IRS operations, the benefits it brings, and the challenges that accompany its adoption.

Detecting Tax Evasion and Fraud

One of the IRS’s most significant applications of AI is in identifying tax evasion and fraudulent activity. By leveraging machine learning algorithms, the IRS analyzes vast datasets of tax returns to detect patterns indicative of noncompliance.

  • Risk-Based Collection Model: The IRS employs AI to prioritize high-risk cases for audits, focusing on wealthy individuals and large partnerships. For instance, AI has been instrumental in targeting 75 of the largest U.S. partnerships, each with assets exceeding $10 billion, including hedge funds, real estate investment partnerships, and law firms. This model has helped recover billions in unpaid taxes by identifying complex tax avoidance schemes.
  • Partnership Audits: With partnerships growing by nearly 600% from 2002 to 2019, AI helps analyze their intricate structures to flag high-risk returns. The Large Partnership Compliance (LPC) program, launched in 2021, uses machine learning to assess accounting rules and tax law compliance, enabling efficient audits of complex entities.
  • Income Matching and Outlier Detection: AI compares taxpayer-reported income with third-party data, such as bank records and property transactions, to issue CP-2000 notices for discrepancies. It also identifies outliers within specific industries, enhancing traditional Discriminant Function (DIF) scores used for audit selection.

Enhancing Audit Selection and Compliance

AI plays a critical role in optimizing the IRS’s audit processes, ensuring resources are directed toward cases with the highest likelihood of noncompliance.

  • Annual Audit Selection: Machine learning models select representative samples of tax returns for audits, prioritizing those likely to have errors or owe additional taxes. This data-driven approach improves audit efficiency and revenue recovery.
  • Earned Income Tax Credit (EITC) Audits: The IRS has piloted AI to select EITC claimants for audits. However, studies have raised concerns about bias, noting that Black taxpayers were audited at higher rates due to algorithmic issues, prompting calls for improved fairness and transparency.
  • Large Partnership Compliance Program: Expanded with AI, this program targets complex partnership returns, using predictive models to identify noncompliance indicators and streamline enforcement efforts.

Improving Taxpayer Services

AI is also enhancing the taxpayer experience by automating routine interactions and improving access to services.

  • Chatbots and Voicebots: AI-powered virtual assistants handle over 450,000 taxpayer inquiries annually, resolving issues without human intervention. These systems use natural language processing to answer questions, provide guidance, and reduce wait times, allowing IRS staff to focus on complex cases.
  • Modernized e-File (MeF) System: AI, including optical character recognition, automates data extraction from paper returns, significantly reducing manual entry and processing backlogs. The MeF system validates returns in real time and offers integrated payment options, improving efficiency and accuracy.

Combating Fraud and Emerging Threats

The IRS Criminal Investigations branch leverages AI to stay ahead of sophisticated fraud schemes and emerging threats.

  • Check Fraud Mitigation: In fiscal year 2023, AI helped recover $375 million by detecting patterns in operational and third-party data to combat check fraud. Machine learning identifies new fraud methodologies, including those enabled by AI-driven deepfakes and cryptocurrency transactions.
  • Foreign Bank Account Reporting (FBAR): AI analyzes multiyear filing patterns to identify potential non-filers with significant foreign account balances, averaging over $1.4 million. This enables targeted audits of high-value cases, strengthening compliance with international tax obligations.

Streamlining Operational Efficiency

AI is transforming internal IRS processes by automating routine tasks and optimizing resource allocation.

  • Automation of Tax Processing: AI streamlines the processing of tax returns, reducing manual data entry and minimizing errors. This allows the IRS to handle large volumes of returns more efficiently.
  • Resource Optimization: By automating simpler cases, AI frees up IRS personnel to focus on high-priority, complex compliance issues, maximizing the impact of limited resources.

Challenges and Concerns

While AI offers significant benefits, its adoption by the IRS has raised several concerns:

  • Bias and Fairness: Algorithmic biases, particularly in EITC audits, have led to disproportionate scrutiny of certain groups, such as Black taxpayers. This has sparked calls for greater transparency and fairness in AI model development.
  • Privacy Concerns: Some critics, including posts on X, have expressed worries about AI accessing financial data without warrants. While these claims lack conclusive evidence, they highlight the need for clear privacy safeguards.
  • Transparency and Accountability: The IRS has faced criticism for not fully documenting AI model specifications or releasing training data, raising questions about accountability and potential inaccuracies.
  • Technological Limitations: Outdated systems hinder full AI integration, and the IRS’s Strategic Operating Plan has not fully aligned with tax gap data, limiting the effectiveness of some initiatives.

Recent Developments and Future Outlook

In March 2025, the IRS paused technology modernization investments to evaluate AI’s role, signaling a strategic shift to leverage the “AI boom” for enhanced collections. Treasury Secretary Scott Bessent has emphasized AI’s potential to improve efficiency without increasing staff, particularly after recent workforce reductions. As the IRS continues to refine its AI applications, it aims to balance innovation with fairness, transparency, and privacy.

Conclusion

The IRS’s use of AI is revolutionizing tax administration by improving fraud detection, audit efficiency, taxpayer services, and operational processes. While these advancements have recovered significant revenue and streamlined operations, challenges like bias, privacy concerns, and transparency must be addressed to ensure equitable and effective implementation. As AI technology evolves, the IRS is poised to further harness its potential to close the tax gap and serve taxpayers more effectively. For more information on IRS initiatives, visit www.irs.gov.

To see past publications please visit our Knowledge Center.

The information presented here should not be construed as legal, tax, accounting, or valuation advice. No one should act on such information without appropriate professional advice and after a thorough examination of the particular situation.