Machine learning enables organizations to detect patterns and emerging trends in their operations. In 2017, the Internal Revenue Service (IRS) adopted machine learning techniques—including data mining and clustering—to detect identify fraud and other types of noncompliance by individuals claiming tax refunds. The Biden administration’s proposal to increase the IRS’s budget by $78 billion over the next decade would use some of those additional funds to further develop machine learning to improve the agency’s targeting of enforcement actions, including audits.

On Wednesday, June 22, the Urban-Brookings Tax Policy Center will host an event examining the potential for using machine learning for tax enforcement. The event will feature a panel discussion among experts in tax enforcement, the use of machine learning to detect noncompliance, and the implications of using artificial intelligence and emerging data technologies on society and governance.

Speakers

• Alex Engler, Fellow, Brookings Institution
• John Guyton, Associate Director, Knowledge, Development, and Application, Internal Revenue Service
• Una-May O’Reilly, Principal Research Scientist and Leader of the Anyscale Learning for All Group, Massachusetts Institute of Technology
• Janet Holtzblatt, Senior Fellow, Urban-Brookings Tax Policy Center (moderator)

Materials

• Speaker biographies
• Primer