The Internal Revenue Service's efforts to leverage artificial intelligence are running into obstacles thanks to the severe cuts in staffing and funding at the agency, according to a new report.
The
The IRS had been ramping up its
The IRS was originally able to leverage the extra $80 billion in funding over 10 years provided by the Biden administration's Inflation Reduction Act of 2022, but those funds were
The IRS has been developing AI applications for many years, but the number grew rapidly until last year, and many of them were considered too sensitive to divulge publicly. The GAO report found the 126 AI "use cases" included 65 that were either too sensitive for public reporting or were research and development efforts exempt from public reporting. "While the IRS has been using AI for several years, its inventory has grown rapidly since reporting 10 use cases in August 2022," said the report.
The IRS has numerous AI initiatives under development and in operation, including in areas such as taxpayer service and audit selection. "However, future IRS funding, strategy and staffing levels are uncertain," said the report.
The IRS categorized most of the use cases in the June 2025 inventory as either for improving operational efficiency or for tax compliance and fraud detection. Some of the applications were finalized and others were still under development last year. The IRS listed 77 of the 126 (61%) use cases as in development in June 2025.
Major staffing reductions at the IRS in 2025 could greatly affect the agency's ability to use AI, the report warned. For example, officials in the IRS's Research, Applied Analytics and Statistics group said they lost 63 employees who had been working either full-time or part-time on AI. Other IRS units also saw reductions in staff that support AI efforts, in addition to organizational and contractual changes.
IRS officials told the GAO the agency plans to use more AI in the future. However, the officials also admitted they hadn't identified the skills needed to support AI nor developed a plan to address the skills gaps.
According to IRS officials, AI expenditures can be hard to track as they include both direct costs that may be traceable to one or more individual use cases and indirect costs that aren't captured at the use case level. Direct costs, including salaries for AI experts, data analysts, and contractors to build an AI model, can be attributed to a specific use case. Contracts can also be used to purchase commercial tools or custom solutions. Indirect costs, such as IT platforms that have enough capacity to process large volumes of data or improve AI performance and reliability, can't be linked to one specific use case, but they're necessary for the IRS to operate AI tools.
"In addition, the IRS funds administrative activities related to its AI efforts, such as governance and procurement oversight to ensure compliance with applicable requirements," said the report.
The IRS has previously estimated that it spends tens of millions annually on AI. For example, in September 2024, the IRS estimated that the two business units most heavily involved with AI — IT and the Research, Applied Analytics and Statistics unit —spent over $58 million on direct and indirect AI costs in fiscal year 2024. In September 2025, RAAS estimated it spent over $28 million on AI in fiscal year 2025 and would spend an additional $32 million in fiscal year 2026.
The IRS reported these estimates included existing contractor, software or hardware investments, but they didn't include costs such as ongoing maintenance for use cases or certain staff costs. According to the IRS, those staff costs could add over 30% to the stated estimates.
"The recent staff reductions, the intent to pursue additional AI initiatives, and the absence of a plan to address AI skills gaps increase the risk that IRS AI efforts will not succeed," said the report.
In addition, the report noted the IRS's inventory didn't always include quality information about the AI applications. For example, the GAO found that over 25% of the use cases lacked information on how they were supposed to benefit the IRS. The GAO also noticed some omissions in the use case inventory. For example, the GAO identified several AI-enabled tools that IRS officials said were contracted to help build criminal cases, but those tools weren't included in the inventory. The GAO believes improved IRS processes and internal communications can address such shortcomings.
The IRS's AI governance process had several entities with oversight of the individual AI use cases, but none of them were responsible for managing AI investments across the agency. The IRS lacks a process to ensure its AI investments are contributing to its agencywide goals.
"Given the risks facing [the] IRS, a more strategic approach is warranted that enables [the] IRS to identify high-value AI initiatives that contribute to agency-wide goals," said the report.
The GAO made eight recommendations in the report to the IRS, including to identify skills gaps and develop an AI workforce plan; implement a comprehensive quality assurance process for AI inventory entries; clarify internal communications to ensure all AI use cases are included in the inventory; and require reporting on use case alignment to strategic goals.
The IRS agreed with all eight of the GAO's recommendations and described the steps it plans to take, or has started taking, in response to each recommendation.
"In alignment with recent executive orders and guidance from the Office of Management and Budget, IRS's uses of AI are growing rapidly," wrote IRS CEO Frank Bisignano in response to the report. "Our focus continues to be on employing AI to improve operational efficiency, tax compliance and fraud detection, and taxpayer services."
Tax pro perspective
Tax professionals have been seeing the IRS using Discriminant Income Function, or DIF, scores for many years to help the agency rate the potential for unreported income on specific tax returns, and AI may help in that area.
"Basically, it's a grade on a tax return, and if the return fails that, then the return would be kicked out for future review," said Tom O'Saben, director of tax content and government relations at the National Association of Tax Professionals, during a press call Tuesday.
"That's what the IRS is using AI for, so that they can avoid the personnel, which they don't have," he added. "They can use AI, along with their DIF scores, to kick out those returns and make the determination."
The AI technology will be able to help sort through the returns to identify ones that could benefit the Treasury.
"That's Big Brother looking over your shoulder and now using technology," said O'Saben. "I really believe that their ability to test returns based on this old risk versus reward standard that took human intervention will now be done by AI. And I think people will see audits coming — a lot of automated ones — where they're projecting some changes or questioning some deductions that were taken. But it will be done with artificial intelligence, which makes the process more efficient. It doesn't replace the individual, because there still will probably be Collections individuals who then will use that, unless it will go to an automated service to do that. But I think that the automated ability to check those returns that failed that DIF score is going to be greatly enhanced just by the existence of artificial intelligence."
