The Impact of Algorithmic Literacy and Purpose on Educator Use and Confidence in Algorithmic Advice

• In Event: Virtual Agents, Artificial Intelligence, and Automation

Abstract

Algorithms may be better at prediction than humans in a variety of contexts, however, they are not perfect. An understanding of the ways educators use algorithmic advice is needed to ensure that accurate algorithmic insights are being utilized, make certain that educators have the knowledge and skills needed to identify poorly constructed or purposely biased algorithms, and attempt to avoid unintended consequences. This quantitative study uses a 2x2 pretest-posttest experimental design in a virtual environment to investigate how educators use and trust algorithmic advice, including the way contextual framing (i.e., algorithm purpose and educational information) and individual-level differences (i.e., teaching experience) impact these relationships. Group comparison and multiple regression techniques will be utilized to test hypotheses.

Author

• Jessa Henderson, The George Washington University
  Presenting Author