My Attorney Is A Robot

In an era where technological advancements are reshaping various sectors, the Internal Revenue Service is not lagging. Thanks to the funding from the Inflation Reduction Act, the IRS is embarking on a transformative journey. Central to this transformation is the adoption of artificial intelligence in its tax audits and compliance operations. This shift towards AI-driven processes marks a significant turning point in the field of tax administration, bringing substantial implications for a wide range of taxpayers, especially those with high incomes and those involved in large business entities.

AI in IRS Tax Audits: The New Frontier

Empowered by the Inflation Reduction Act (IRA, P.L. 117-169), the IRS is poised for a significant technological upgrade, with a special emphasis on artificial intelligence. This move builds upon previous initiatives where AI, including machine learning, natural language processing, and graph analytics, was used to enhance the detection of tax fraud and other illegal activities. These advancements form a critical part of a wider strategy aimed at strengthening tax enforcement, particularly against complex financial entities.

At the heart of this new phase in AI integration is a two-fold strategy: system upgrades and the strategic recruitment of data scientists. These experts will play a pivotal role in developing advanced AI tools within the organization. This approach will arm the IRS with state-of-the-art technology, enhancing its capacity to address increasingly complex tax challenges.

Benefits of the Utilizing of AI

The integration of AI into the IRS’s operations has reportedly already resulted in improved efficiency and effectiveness. For example, using chatbots for customer service has streamlined taxpayer communications, resulting in more than 450,000 interactions, 42% of which were resolved without needing live agents. This not only spared taxpayers the inconvenience of waiting on hold but also allowed agents to devote their time to other tasks.

In the fraud detection realm, AI has also proven highly effective. For instance, the IRS reported that in 2019, its AI technology identified 296 suspected fraudulent returns, totaling nearly $1.3 million—a significant achievement considering that 84% of these were not flagged during manual reviews.

Looking ahead, the use of AI is expected to become even more prevalent. It will likely play a crucial role in analyzing complex tax returns, especially those from large partnerships. AI’s ability to meticulously analyze extensive documentation and interlink data across different entities makes it an invaluable tool for uncovering noncompliance signs that might otherwise remain hidden.

Concerns With the Utilization of AI

The implementation of AI in tax enforcement, while beneficial, also raises significant concerns about accountability. This is particularly relevant in terms of civil liberties, ethics, and social disparities, as pointed out by the Government Accountability Office. The primary issue lies in the inherent lack of transparency and the potential for biases within AI systems, which risk perpetuating historical data processing problems. An example provided by Abdi Aidid of the University of Toronto Faculty of Law at the ABA meeting on May 4 highlights this concern: AI tools used for setting optimal wages could reinforce existing gender wage disparities if they fail to account for factors like women’s more frequent career interruptions and historically lower wages.

In 2021, a real-world example in the Netherlands vividly illustrated the dangers of unaccountable AI use. There, an AI tool designed to identify fraud in child care subsidies was discovered to have unjustly targeted applicants with foreign origins and demanded repayment of benefits. This issue was so severe that it led to the collective resignation of the entire Dutch government.

Similarly, AI-driven decision-making at the IRS poses significant risks. For instance, the IRS’s use of algorithms, often opaque, can lead to decisions that are difficult to understand or contest. This is compounded by the IRS’s policy of keeping its algorithms confidential, leaving taxpayers unaware of whether their cases were decided by these systems without effective human oversight. An alarming instance of this was reported by the Treasury Inspector General for Tax Administration in 2012. TIGTA found that an IRS algorithm incorrectly decided 89% of cases involving penalty relief, a mistake that was never corrected by human oversight.

This issue extends to AI algorithms potentially targeting specific taxpayer demographics. As The Hill contributor Robert J. Kovacev pointed out, an IRS algorithm might conclude from collection data that low-income taxpayers are less likely to contest tax assessments, leading to a disproportionate increase in audits of these taxpayers. In fact, according to a 2021 analysis of internal IRS reports, low-income families are already audited at five times the rate of other taxpayers. The inscrutable nature of these algorithms means that such biased targeting could go unnoticed.

Potential Targets of IRS AI Technology in the Near Future

Regardless of the pros and cons, it is anticipated that the IRS will expand its use of AI enforcement into a few key areas in the near future. One primary target will be complex financial transactions, particularly those involving international transfers and offshore accounts. The global nature of modern finance presents unique challenges, and AI’s advanced analytical capabilities are well-suited to unravel complex webs of transactions that might otherwise evade traditional scrutiny. This could mean increased scrutiny of multinational corporations and high-net-worth individuals with assets and accounts spread across multiple jurisdictions.

Another area of focus will likely be the burgeoning field of cryptocurrency and digital assets. The IRS has already begun to pay closer attention to these transactions, but with the aid of AI, it could significantly ramp up enforcement. AI can analyze patterns in blockchain transactions to identify potential tax evasion or fraud. This would not only apply to large-scale investors and traders but also to smaller individual participants in the cryptocurrency market.

The IRS is also expected to use AI to enhance its auditing capabilities for large partnerships and corporations. The complexity of these entities’ tax returns, involving layers of partnerships, deductions, and credits, makes them ripe for AI analysis. The goal here would be to identify inconsistencies or anomalous reporting that might indicate underreported income or overclaimed deductions.

Additionally, the agency’s focus on data-driven decision-making suggests that it might also target areas where it has vast amounts of data but has historically had low enforcement, such as small businesses and sole proprietorships. AI can sift through large datasets to identify patterns of noncompliance that might have been overlooked in the past.

In all these areas, the key advantage of AI is its ability to process and analyze data at a scale and speed unattainable by human auditors. This capability, combined with the IRS’s increasing sophistication in using these tools, means that taxpayers in these categories may face a higher likelihood of audits and investigations in the near future.

Conclusion

The IRS’s integration of artificial intelligence into its tax administration processes marks a significant turning point, though it’s too early to fully gauge the impact of AI on identifying tax evasion or patterns of noncompliance. This move into a more technologically advanced era is momentous but comes with substantial responsibilities.

Key among these responsibilities is addressing the crucial issues of transparency, accountability, and potential biases within AI systems. It’s essential for the IRS to ensure that AI is applied fairly and doesn’t unjustly target compliant taxpayers, especially in complex areas like international transactions and cryptocurrency. As the agency shifts to a more data-driven and analytical approach, it must strike a delicate balance between embracing technological innovation and protecting civil liberties and taxpayer rights.

Moving forward into this era of AI-enhanced tax enforcement, the overarching goal must be to advance without compromising fairness and trust. The IRS faces the dual task of leveraging AI’s potential while upholding ethical standards and treating all taxpayers equitably. Achieving this balance is critical not only for modernizing the IRS’s operations but also for establishing a tax system that is fair, just, and responsive to the evolving needs and expectations of society.