AI’s Potential to Increase IRS Audits Raises Expectations and Fears

Key Takeaways

• IRS spent $1.8 M on Palantir’s SNAP AI audit tool
• SNAP consolidates data from 100+ legacy systems into one platform
• Audit rate 0.3%; AI may raise it for high earners
• Effective use depends on retaining skilled staff for human review

Pulse Analysis

The Internal Revenue Service’s recent partnership with Palantir Technologies marks a significant shift toward artificial‑intelligence‑driven compliance. The $1.8 million SNAP (Selection and Analytic Platform) contract aims to unify a sprawling network of over a hundred legacy databases, allowing the agency to apply machine‑learning models that flag "highest‑value" returns for deeper examination. By moving beyond the decades‑old Discriminant Information Function, which relied on static formulas, SNAP promises to detect nuanced patterns of under‑reporting, credit abuse, and potential fraud that human analysts might miss.

For taxpayers, the prospect of AI‑guided audits raises both alarm and curiosity. The IRS’s latest Data Book shows a modest 0.3 percent audit rate for individual returns in 2019, but the odds climb sharply for high‑income earners—reaching 1.6 percent for those earning $1‑5 million and over 11 percent for incomes above $10 million. SNAP could amplify these disparities by pinpointing sophisticated avoidance schemes, potentially increasing audit exposure for the affluent while leaving compliant filers untouched. This dynamic underscores a broader debate about fairness, transparency, and the balance between revenue collection and taxpayer rights.

Implementation, however, is far from automatic. Critics warn that the IRS’s shrinking workforce—exacerbated by budget cuts and efficiency drives—may limit the human expertise needed to interpret AI outputs reliably. Without adequate reviewer capacity, false positives could surge, eroding public trust. Moreover, legislative scrutiny over AI’s role in government surveillance adds a layer of regulatory risk. As the pilot progresses, policymakers and the agency must grapple with ensuring robust oversight, safeguarding data privacy, and establishing clear accountability frameworks to harness AI’s benefits without compromising the integrity of the tax system.