IRS Boosts AI for Audits Amid Staffing, Funding Shifts

TL/DR –

The Internal Revenue Service (IRS) is facing staff losses, budget shifts and increased reliance on artificial intelligence (AI) for auditing, which is creating uncertainty for taxpayers. The IRS has seen a rise in staff resignations, particularly among those involved in enforcement and complex audits, while also undergoing funding cuts. At the same time, the agency is modernizing its systems and incorporating AI to assist in identifying noncompliance and fraud, but concerns are being raised over whether there will be enough experienced staff to handle potentially suspicious cases identified by this technology.

The IRS and AI: Examining the Future of Audits

With recent decrease in staff and shifting budget priorities at the IRS, tax audits have become a major point of discussion. Concerns about who will be audited and how the enforcement will work have been raised, especially given the rapid adoption of AI technologies.

The State of Audits

The frequency of audits has been quite low in recent years. According to the numbers, only 0.3% of tax filers experienced an audit for the 2021 tax year. The percentage of those audited varies by income level, sources of income, and claimed deductions, but even in these varying groups, less than 10% faced an audit, with most groups seeing rates far below 1%.

Shaking Up the IRS

Change has been a constant at the IRS over the past year. An upward trend has been noted in the number of experienced professionals in complex audits and enforcement leaving the agency. As stated in the July 2025 report from the Treasury Inspector General for Tax Administration, more than a quarter of tax examiners and revenue agents exited their positions.

Additionally, a significant amount of planned enforcement funding under the Inflation Reduction Act of 2022 has been canceled. This comes as the Trump administration sets its sights on reducing IRS funding even further in the coming year.

Concurrently, the IRS is making strides to update its aging systems and incorporate artificial intelligence into numerous areas, including enforcement. As Frank Bisignano explains, the “IRS uses artificial intelligence (AI) and advanced analytics to more accurately identify high-risk areas of noncompliance and fraud.”

One of the primary objectives of these changes is to strengthen the compliance program and exceed the usual revenue expectations.

The Effect on Audit Probability

How these changes will impact the likelihood of audits remains uncertain. Many questions have been raised about the strategic and responsible use of technology, and whether there will be enough experienced staff to properly audit returns identified as potentially suspicious by the AI systems.

The goal with AI is to improve efficiency and effectiveness in the audit process. It helps the IRS to avoid “audits with no changes” — audits that yield no additional revenue due to a lack of violations. Additionally, AI is adept at detecting underreported income and other tax rule violations.

Danny Werfel notes that “AI gives us a forensic edge in selecting the right returns for audit.” AI’s ability to identify patterns in tax returns that may otherwise go unnoticed can considerably reduce the time necessary to detect anomalies indicating miscalculations or flagrant fraud.

However, it is emphasized that AI should not replace human oversight. Monitoring and validating the decisions made by AI will help maintain the reliability of the results. This includes emphasis on increasing the share of “correspondence” audits, which are more cost-effective than field or office audits, and involve sending the taxpayer a notice of possible additional tax or withholding a refund for certain items.

Implementing AI in this way requires a sufficient number of qualified professionals to respond to taxpayer inquiries, especially when AI identifies issues with more complex returns that require field or office audits. But with the recent decrease in staff, there are concerns about the availability of experienced enforcement personnel in the job market.

“I don’t think AI will replace the staff that was lost. You still need experts who can meet with filers and check their books and records to determine whether your return is correct,” Johnson, an independent fellow at the Urban-Brookings Tax Policy Center, added.

As AI evolves, it may offer even more capabilities in the future. As Danny Werfel warns, “When you decide to be aggressive, bend the rules, or break them, you should look at the IRS’s potential not only now, but in the future.”

The IRS generally has three years from the date a return is filed to audit it, but this period can be extended under certain circumstances, such as suspicions of fraud. Therefore, the dynamics of future audits will depend on the blend of technology and human oversight. While AI can make the process more efficient and accurate, the final decision will still lie with the professionals.

—
Read More US Economic News