The IRS is using artificial intelligence (AI) to help select tax returns to audit as it tries to narrow the tax gap, Accounting Today reported.

The tax gap refers to the difference between what is owed and what is paid by taxpayers. Taxpayers voluntarily paid 85 percent of taxes they owed on time in 2014-2016, but the other 15 percent of taxes owed went unpaid, much of that due to underreporting, the Government Accountability Office stated in a report on the IRS initiative. That amounted to an estimated $496 billion gap.

“IRS most recently projected the tax gap will grow to $688 billion for tax year 2021,” stated the report. “However, when measured relative to the overall economy, the tax gap has remained relatively stable.”

To address the tax gap, the IRS must use this new technology consistently and transparently, the GAO cautioned.

"To estimate part of the tax gap, the IRS audits a random sample of tax returns based on certain characteristics—such as whether the taxpayer is self-employed," said the report. "The agency has recently piloted a new AI method to select taxpayers for audits. However, it hasn't documented key components and technical specifications of this AI. We recommended that it do so to ensure that the IRS is being consistent and transparent with its AI use."

The GAO made six recommendations to the IRS in the report, suggesting the IRS should conduct additional analyses to understand the root causes of undetected noncompliance and complete its documentation related to the pilot process for sampling returns. Specifically, it recommended that the IRS commissioner should:

1) ensure officials in the Research, Applied Analytics and Statistics Division work with officials in the Small Business/Self-Employed and Large Business and International Divisions to develop and report on analyses that could provide stakeholders with more confidence in the methods used to estimate undetected noncompliance, and information on its root causes;

2) complete documentation on key components of the National Research Program (NRP) AI models, such as which model options were considered and chosen for implementation, and how IRS determined risk levels and time frames for risk level updates;

3) complete documentation for the NRP AI models on the technical specifications needed to run the model, such as how the data are to be split or divided;

4) ensure that the IRS has complete documentation on how to update the NRP sample selection process in response to changes in the operating environment, including the performance of its AI models and risk level analysis;

5) complete documentation on processes for assessing the results of AI model updates, including validating the models; and

6)  develop and document a plan for evaluating the redesigned sample selection process, including factors IRS will consider in determining whether or how to continue using it.

The IRS agreed with all six of the GAO's recommendations and described the steps it plans to take in response to each recommendation.

"We agree that there is always room for improving existing technical documentation to make it more accessible, and we recognize the importance of formally documenting analyst and management decisions," wrote IRS chief operating officer Melanie Krause in response to the GAO report, Accounting Today reported. "We will incorporate your recommendations moving forward."

Accounting Today noted that the IRS has been developing AI technology for several years. "I'm a big fan of artificial intelligence and machine learning," said Eric Hylton, a former commissioner of the IRS's Small Business/Self-Employed (SB/SE} division and deputy chief of the IRS's Criminal Investigation (CI) division who is now national director of compliance at the tax consulting firm Alliantgroup, in a recent interview with Accounting Today. "There are definitely some areas that need improvement. In any particular system, 'garbage in, garbage out,' so you have to make sure that you're actually doing quality control over the information that you actually put into your system. However, it allows the service to be a lot more efficient. You can have analytics related to areas of noncompliance."

He noted that when he was SB/SE commissioner working with CI, one of the things that SB/SE would like to do before sending a referral to CI would be to analyze, based on previous referrals, which ones actually led to a successful prosecution.

"You can use artificial intelligence and predictive analytics to be able to analyze that input and take those factors, and then place them over your current inventory and see if there are cases that you should be recommending for referrals to Criminal Investigation," said Hylton.