It has been revealed that the IRS plans to implement artificial intelligence (AI) in its audit selection process, targeting high-income individuals, partnerships, and corporations. The announcement came as part of a broader initiative, funded by the Inflation Reduction Act, focusing on numerous subjects including cryptocurrencies, foreign bank account reporting (FBAR) compliance, and deceptive practices, among others. As part of this initiative, large law firms have been singled out for careful examination.
By this month’s end, the IRS intends to commence audits of 75 of America’s largest partnerships, spanning a range of industries such as hedge funds, real estate investment partnerships, and publicly traded partnerships, not to mention large law firms. Typically, these partnerships possess more than $10 billion in assets each.
The IRS also envisages reaching out to 500 partnerships via mail in early October. These partners have an imbalance on their sheets exceeding $10 million in assets, suggestive of possible non-compliance with the IRS. The nature of the audit selection will be informed by AI.
IRS’s announcement states, “With the help of AI, the selection of these returns is the result of collaboration among experts in data science and tax enforcement, who have been working to apply machine learning technology to identify potential compliance risk in the areas of partnership tax, general income tax and accounting, and international tax in a taxpayer segment that historically has been subject to limited examination coverage.”
Despite the opacity of the IRS on this aforementioned ‘collaboration’ or what precisely ‘machine learning technology’ is, the selection methodology has historically been confidential. This includes the referencing of a “discriminant function” score, often regarded as a “DIF score” to discern the potential of tax returns for audit. This score, however, is not something taxpayers are privy to.
The reason behind the IRS targeting these large partnerships stems from the popularity of such business formats among high-income industries like real estate and law firms. This stems from the tax law’s flexibility in allocating gains, losses, deductions, and accreditation. However, this flexibility may also lead to misuse. The law has anti-abuse regulations to mitigate this, requiring that these allocations have a “substantial economic effect”.
The complexity of partnership tax law, oftentimes deemed convoluted, may lead to unintentional noncompliance, such as in the case with the Treasury regulation detailing “substantial economic effect”. This document, known as 704(b) regulations, spans 144 pages.
A 2022 report from the Treasury Inspector General for Tax Administration revealed that audits of partnership returns are not as successful as expected, exhibiting a 78% no-change rate with audits of 480 partnership returns between 2016 and 2019. This is compared to a 50% no-change rate for all partnership returns during the same timeframe.
Large law firms, being likely targets of the IRS’s attentions, should reevaluate their previous partnership returns. It is important to note that the IRS has a period of three years from filing date to conduct an audit. However, in instances of “substantial understatement of income” – where the return omits over 25% of gross income – the IRS has a six-year window. And, with fraudulent returns, there are no time constraints for audits.
Originally published on Above the Law.