The Treasury Inspector General for Tax Administration identified 32 potential violations of Fair Tax Collection Practices by IRS employees between July 2024 and June 2025, according to a semiannual audit required by Congress.
Most violations involved contacting a represented taxpayer or bypassing an authorized representative. Of the 32 identified violations, 16 were correctly coded as Fair Tax Collection Practices violations in the IRS's tracking system, while the remaining 16 were incorrectly coded as other types of misconduct and discovered through additional research. The Treasury Inspector General had previously recommended enhanced training to address such miscoding.
The IRS is also facing scrutiny over artificial intelligence risk management. An audit examined the IRS's documentation of risk assessments for high-impact AI use cases, as required by an Office of Management and Budget memorandum. The IRS has used AI systems for many years, including for tax return classification and issue selection, with 225 AI use cases documented as of December 2025.
The audit reviewed five presumed high-impact AI use cases and found deficiencies in two areas. Two of the five cases, or 40 percent, lacked documented impact assessments. For data quality assessments, only one of the five cases had comprehensive documentation of testing procedures, while four cases lacked testing documentation. The IRS had not established standardized procedures for evaluating data quality before using data in high-impact AI models.
Treasury had not issued guidance on minimum risk management practices as of July 2026, though the IRS has developed its own AI governance policy to comply with an Office of Management and Budget requirement for implementation by April 2026.