AI Symposium: Final Panel Details, and Third Panel Recap

The fourth and final installment of ACUS’s Symposium on Artificial Intelligence in Federal Agencies will take place on Thursday August 13, from 1:00 p.m. to 2:15 p.m. At the panel, titled Artificial Intelligence In The Trenches: A View From Inside The Agencies, four current and former federal agency officials will discuss how their agencies currently use or contemplate using AI to support their missions and improve the government’s efficiency and effectiveness. Registration for the event is available here

Last week ACUS, in partnership with the Institute for Technology Law & Policy at Georgetown University Law Center, hosted the third panel in the series, titled Bias and Government Artificial Intelligence.

Moderated by ACUS Public Member, Morgan Lewis Partner, and former EEOC Commissioner Chai Feldblum, the panel included Tulane University Law School Professor Kristin Johnson, Georgetown University Law Center Professor David Super, and Center for Democracy & Technology President and CEO Alexandra Givens.

The panelists described several instances where the federal government’s use of AI has had a disparate or unintended impact on certain communities, as well as instances where there may be a role for government agencies to regulate private-sector uses of AI with similar effects. Professor Johnson discussed how private-sector financial companies’ unregulated use of AI systems to determine creditworthiness can have a disparate impact on minority groups. Professor Super explained that the government’s AI may misidentify fraud in the Supplemental Nutrition Assistance Program (SNAP), and many lawyers may lack the tools or training to identify errors in decisions rendered by black-box AI systems against their clients. Ms. Givens addressed how the increased automation of tasks essential to determining eligibility for state-administered public benefits has resulted in serious errors and litigation.

Following their initial presentations, the panelists discussed issues such as the advantages and disadvantages of centralizing AI expertise in the federal government and the role of agency attorneys to ensure that government AI systems comply with legal norms and policy objectives.

Previous panels in the series featured the authors of a recent, ACUS-commissioned report, Government by Algorithm: Artificial Intelligence in Federal Administrative Agencies, and scholars studying the intersection of AI and core administrative and constitutional law doctrines. Recordings of the first three panels are available at

ACUS looks forward to the continued exploration of AI in the administrative state, including a forthcoming report from Professor Cary Coglianese and the continued compilation of resources on our website.

Share on FacebookShare on TwitterShare on Linkedin