HR-9501-119
Ordered to be Reported in the Nature of a Substitute by the Yeas and Nays: 40 - 0.
Sponsored by Vern Buchanan (R-FL)
What it does
This bill would require the Treasury Department to launch an artificial intelligence pilot program within 180 days of enactment to identify inaccurate or fraudulent tax returns, including those involving identity theft, false credit or refund claims, and returns improperly prepared by unidentified third parties. The pilot program would run for 18 months to 2 years. After it ends, the Comptroller General (head of the Government Accountability Office) would have 180 days to submit a report to the House Ways and Means Committee and Senate Finance Committee detailing how much fraud was detected, how much money was recovered, and how accurate the AI tools were.
Who benefits
Honest taxpayers who lose refunds or have their identities stolen by fraudulent filers. The U.S. Treasury, which may recover improperly paid refunds. Tax compliance professionals and legitimate tax preparers, who face less competition from fraudulent preparers. Congress, which would receive structured data to inform future AI and IRS policy. AI technology vendors who may be contracted to build or operate the pilot program.
Who is hurt
Taxpayers whose returns are flagged as potentially fraudulent by AI tools — including those who are flagged incorrectly — may face delays in receiving refunds or additional IRS scrutiny. Individuals who file complex but legitimate returns may face a higher rate of false-positive flags. Tax preparers who operate in gray areas could face increased detection risk. Civil liberties and taxpayer privacy advocates may raise concerns about AI access to sensitive financial data.
Supporters argue
Supporters argue that the IRS loses an estimated $5–$7 billion annually to fraudulent refund claims and that existing detection methods have not kept pace with increasingly sophisticated fraud schemes, including AI-generated false returns. They contend that a time-limited, GAO-evaluated pilot program is a measured, evidence-based approach that allows Congress to assess AI's effectiveness before committing to broader deployment, protecting both taxpayer dollars and civil liberties through built-in oversight.
Opponents argue
Opponents argue that deploying AI to screen tax returns raises serious privacy and due process concerns, as algorithmic systems can produce biased or inaccurate results that disproportionately burden lower-income filers and communities of color who already face higher audit rates. They contend that without explicit accuracy thresholds, appeal rights, or anti-bias safeguards written into the bill, the pilot program could normalize AI-driven enforcement with inadequate protections for taxpayers who are incorrectly flagged.