What are your chances of being audited now that the IRS is using AI? Jury is still out

The IRS audited just 0.3% of filers in tax year 2021, but the agency is now using AI and advanced analytics to better identify noncompliance and fraud. The article argues AI could increase correspondence audits, though staffing cuts, lost enforcement expertise, and evolving AI capabilities make the eventual impact on audit rates unclear. The agency generally has at least three years to audit a return, longer in some fraud cases.

Analysis

The market implication is not “more audits,” but a reallocation of enforcement from broad-based friction toward highly selective, potentially higher-conviction collections. That tends to favor taxpayers with clean, standardized reporting and hurt edge cases where complexity, pass-through structures, or high-variance deductions create pattern noise; the second-order winner is any software or service stack that helps taxpayers pre-screen filings and document positions before submission.

The near-term constraint is execution capacity, not model quality. If AI raises issue-spotting faster than the IRS can staff correspondence handling and complex examinations, the bottleneck shifts from detection to follow-through, which means headline audit rates may not move much even if collection intensity rises in specific cohorts. That creates a base case of uneven enforcement: more letters, fewer fully prosecuted audits, and a higher probability of delayed adjustments rather than immediate revenue realization.

The contrarian point is that the biggest risk is not a sudden jump in audits this filing season, but a lagged improvement in targeting over the next 12–36 months as models learn and staffing normalizes. That makes complacency around “low historical audit rates” dangerous for aggressive filers, while also suggesting the current market may be underpricing compliance-adjacent beneficiaries and overpricing any short-term fear premium in tax-sensitive small caps. The real catalyst is not AI announcements, but evidence of sustained collection conversion and reduced no-change audit waste.

Key Decisions for Investors

• Long H&R Block (HRB) / Intuit (INTU) into next filing season: the setup is cleaner self-serve compliance demand and higher willingness to pay for audit-defense/accuracy tools; risk/reward improves if IRS enforcement messaging remains elevated over the next 6-12 months.
• Consider a small basket long in tax-documentation and workflow software names exposed to SMBs, funded by shorting low-quality tax-prep/filing adjacent discretionary spenders; thesis is that complexity premium rises as enforcement becomes more selective.
• Avoid or reduce exposure to highly aggressive, tax-optimized small-cap structures and flow-through-heavy names over a 12-24 month horizon; these are the most likely to face delayed enforcement surprises if AI improves issue selection.
• For event-driven hedging, buy 12-24 month put spreads on names where reported tax benefits or uncertain tax positions are a meaningful valuation support, as the risk is not immediate audit headlines but multi-quarter reserve build and cash tax leakage.