I.R.S. Deploys Artificial Intelligence to Catch Tax Evasion

The Internal Revenue Service has started using artificial intelligence to investigate tax evasion at multibillion-dollar partnerships as it looks for ways to better police hedge funds, private equity groups, real estate investors and large law firms.

The announcement on Friday was intended to show how a more muscular I.R.S. is using some of the $80 billion allocated through last year’s Inflation Reduction Act to target the wealthiest Americans and tackle the kinds of cases that had become too complex and cumbersome for the beleaguered agency to handle.

The agency’s new funding is supposed to help the I.R.S. raise more federal revenue by cracking down on tax cheats and others who use sophisticated accounting maneuvers to avoid paying what they owe. But the allocation has been politically contentious, with Republicans claiming that the I.R.S. will use the funding to harass small businesses and middle-class taxpayers. Earlier this year, Republicans succeeded in clawing back $20 billion as part of an agreement to raise the nation’s borrowing cap.

The political fight has put the onus on Democrats and the Biden administration to show that the funding is primarily enabling the I.R.S. to target wealthy Americans and corporations who may have engaged in tax evasion.

“These are complex cases for I.R.S. teams to unpack,” Daniel Werfel, the I.R.S. commissioner, said in a briefing with reporters. “The I.R.S. has simply not had enough resources or staffing to address partnerships; in a real sense, we’ve been overwhelmed in this area for years.”

The fight over I.R.S. funding is continuing, as the House and the Senate try to agree on spending legislation to avert a possible government shutdown at the end of the month. Senate Democrats want to hold the base budget of the I.R.S. steady while holding on to some of the Inflation Reduction Act money that lawmakers had agreed to rescind as part of the debt limit deal, while House Republicans are pushing for far deeper cuts that would eat into the tax agency’s enforcement budget.

Senator Ron Wyden of Oregon, the Democratic chairman of the Senate Finance Committee, said the new enforcement tools that the I.R.S. is deploying show the importance of the funding.

“This news stands in stark contrast to the approach taken by House Republicans, who want to allow wealthy tax cheats to continue business as usual, paying little to no tax and asking middle-class taxpayers to foot the bill,” Mr. Wyden said.

Mr. Werfel explained that artificial intelligence is helping the I.R.S. identify patterns and trends, giving the agency greater confidence that it can find where larger partnerships are shielding income. That is leading to the kinds of major audits that the I.R.S. might not have previously tackled.

The agency said it would open examinations into 75 of the nation’s largest partnerships, which were identified with the help of artificial intelligence, by the end of the month. The partnerships all have more than $10 billion in assets and will receive audit notices in the coming weeks.

More audits are likely to come. In October, the I.R.S. will send 500 notifications, known as compliance alerts, to other large partnerships indicating that the agency has found discrepancies in their balance sheets. These partnerships could also face audits if they cannot explain the differences in their balances from the end of one year to the start of the next.

The focus on partnerships is part of a broader push by the I.R.S. to scrutinize wealthier taxpayers in 2024. Mr. Werfel said that the agency is dedicating dozens of revenue officers to pursue 1,600 millionaires who the I.R.S. believes owe at least $250,000 in unpaid taxes.

In the coming year, the I.R.S. said it plans to increase scrutiny of digital assets as a vehicle for tax evasion and investigate how high-income taxpayers are using foreign bank accounts to avoid disclosing their financial information.

The use of artificial intelligence at the I.R.S. opened it up to fresh criticism from those who complain that it cannot be trusted with taxpayer data.

Grover Norquist, founder and president of Americans for Tax Reform, said that the I.R.S. has a history of blaming its problems with enforcing the tax code on its algorithms. He argued that employing artificial intelligence was just another way for the I.R.S. to separate itself from future accusations of political bias or inequitable enforcement practices.

“This is one more way for them to put some distance in their decision-making,” Mr. Norquist said. “They can say, ‘Oh, we’re not auditing people we don’t like. This is science.’”

The I.R.S. has said little so far about how it intends to use artificial intelligence to crack down on tax evasion. Mr. Werfel suggested that the technology would be deployed to identify “compliance threats” that have been difficult to spot and that it would help the agency reduce unnecessary audits.

John Koskinen, a former I.R.S. commissioner during the Obama and Trump administrations, noted that the tax returns of big partnerships are among the most complex that the agency analyzes. He said algorithms that the I.R.S. already uses allow it to spot problems in tax returns, but that the ability to train computers to sift through thousands of pages of tax returns that have red flags would make it possible for the agency to more effectively determine whom to audit.

“Up until now, those partnerships have been impenetrable,” Mr. Koskinen said.

A Government Accountability Office report that was published in July found that the I.R.S. audit rate for larger partnerships had dropped below 1 percent since 2007 and that 80 percent of the audits that the agency did conduct did not find any noncompliance problems.

“As part of its audit selection process, I.R.S. uses statistical models to help review partnership returns for potential noncompliance, but the models were developed without using representative samples of returns and with untested assumptions,” the G.A.O. said.

The I.R.S. has also lacked the ability to compare the tax information of large partnerships to other partnerships to look for additional warning signs, but artificial intelligence is starting to make that possible and is expected to improve over time.

As part of its recruiting strategy, the I.R.S. has been looking to hire data scientists to develop new in-house artificial intelligence tools. Mr. Werfel said that the agency is also collaborating with outside experts and contractors on the project.

Last month, the I.R.S. said it had increased its full-time staff to nearly 90,000, a level not seen in more than a decade.

Mr. Werfel warned that additional cuts to its annual budgets would require him to redirect money meant for upgrading the agency so that it can carry out basic tasks, and that ultimately taxpayers would bear the brunt of it.

“My big responsibility here is to make the case to Congress and to the American people that if you fund our base budget, it will enable us to both keep the lights on and modernize — and modernizing the I.R.S. is good for everybody,” he said.