Sparking a Debate on the Role of Artificial Intelligence in Educational Measurement

Session Type: Organized Discussion

Abstract

Artificial intelligence (AI) has proven a useful and powerful technology for applications across numerous disciplines and industries. The growth of AI tools and applications reflects the zeitgeist of our times—AI is increasingly deployed to gain efficiencies through automation and provide precise predictions. Accordingly, educational researchers, methodologists, and psychometricians are increasingly interested in the role of artificial intelligence (AI) in educational measurement. The practice of educational and psychological measurement is guided by principles and standards. The proliferation of AI has in many ways outpaced The Standards and the purpose of this symposium is to provide a forum for debate concerning the role of AI for educational measurement. Researchers at Duolingo recently were the first educational testing company to produce “Responsible AI Standards” (RAIS) for AI in educational assessment (Burstein, 2023). Through short presentations and moderated debate, several educational measurement leaders will attempt to achieve consensus or draw clear lines of disagreement about how or whether to use AI in educational measurement.

Sub Unit

• Individual Papers, Innovation Demonstrations, Coordinated Sessions, or Organized Discussions

Session Organizer

• Steven Culpepper, UNIVERSITY OF ILLINOIS AT URBA

Moderator

• Brian Junker, CARNEGIE MELLON UNIVERSITY

Presenters

• Alina A. von Davier, Duolingo

• Andrew Ho, Harvard Graduate School of Education

• Ezekiel Dixon-Roman, UNIVERSITY OF PENNSYLVANIA

• Hua-Hua Chang, Purdue University

• Derek Christian Briggs, University of Colorado Boulder

• Matthew S Johnson, ETS