The New Era of IRS Audits: High-Tech Eyes, Fewer Boots on the Ground
For years, the statistical likelihood of facing an IRS audit has been remarkably low. In tax year 2021, for instance, the agency pursued a mere 0.3% of filers overall. Even as certain groups—defined by income levels or specific tax breaks—face higher rates, these typically remain under 10% and often fall below the 1% mark.
However, the landscape of federal tax enforcement is shifting. The IRS is currently navigating a paradoxical transition: it is deploying powerful new technology while simultaneously facing a significant drain of human expertise and funding.
How AI is Changing the Audit Game
The IRS is moving away from antiquated systems and embracing artificial intelligence (AI) and advanced analytics. According to IRS CEO Frank Bisignano, the agency is utilizing these tools to identify high-risk areas of fraud and non-compliance with significantly greater accuracy.
The “Night-Vision” Effect
Former IRS Commissioner Danny Werfel describes the integration of AI as akin to “purchasing night-vision goggles.” Rather than relying on older statistical models, AI can scan massive datasets to identify anomalies and patterns that would be nearly impossible for humans to detect manually.
This “forensic edge” allows the agency to pinpoint returns that may indicate underreported income or outright fraud, potentially increasing the efficiency of how non-compliance is flagged.
The Rise of Correspondence Audits
As AI can isolate discrete issues on a return, the agency may increase its use of correspondence audits. Unlike resource-intensive field or office audits, a correspondence audit is handled via mail.
In these cases, the IRS sends a letter indicating that more money may be owed or that a refund is being withheld based on specific issues. The filer must then either submit payment or challenge the assessment.
The Human Element: The Staffing Paradox
While AI can identify a suspect return, it cannot conduct a full-scale audit. This is where the IRS faces its greatest challenge. A July 2025 report from the Treasury Inspector General for Tax Administration revealed a significant loss of talent, with over a quarter of both revenue agents and tax examiners leaving the agency.
This loss of experienced personnel creates a bottleneck. As Barry Johnson, former chief data and analytics officer at the IRS, notes, AI cannot substitute for experts who meet with taxpayers and review physical books and records to determine accuracy.
Funding Cuts and Future Uncertainty
The agency’s technological ambitions are clashing with financial realities. Much of the enforcement funding promised under the 2022 Inflation Reduction Act has been rescinded and the Trump administration is seeking further cuts to IRS funding in the coming year.
This creates an uncertain future for taxpayers. The ability of the IRS to act on the leads generated by AI depends entirely on whether they have enough trained staff to handle the resulting inquiries and complex reviews. If used responsibly—balancing AI output with human double-checking—the goal is to reduce “no change” audits, where the IRS spends resources only to find no revenue was owed.
As AI continues to evolve, the surveillance and assessment powers of the IRS may shift again by 2028 or 2029, representing a true paradigm shift in how tax evasion is detected.
Frequently Asked Questions
What is a “no change” audit?
A “no change” audit occurs when the IRS conducts an audit but finds no errors or underpayments, resulting in no additional revenue being raised.
How long does the IRS have to audit my taxes?
The agency typically has at least three years from the date you file to conduct an audit, although this period can be longer if fraud is suspected.
Will AI replace IRS agents?
No. While AI can identify high-risk returns and patterns, human experts are still required to review books and records and meet with taxpayers to verify accuracy.
What is a correspondence audit?
It is the most common type of audit, where the IRS communicates via letters regarding specific issues on a return rather than conducting an in-person field or office audit.
What are your thoughts on the IRS using AI to flag tax returns? Do you consider it will make the system fairer or more intrusive? Let us know in the comments below or subscribe to our newsletter for more updates on tax trends.
