Student Detection, Preferences, and Reactions to AI and Human Grading in Educational Contexts
Steven R Wong, Marcelina Leon Zavala
Department of Psychology
Faculty Supervisor: Diana R. Sanchez
Two experimental studies were conducted. The first study examined students' perceptions of using AI tools for schoolwork, assessed which type and style of AI text is most easily identified as AI, which type and style of text is easiest to read, and which is considered higher quality, along with overall confidence in their judgment. The second study explored students' perceptions of instructors' use of AI tools for grading. Students completed a short writing task and were asked about their opinions and perceptions of distributive justice regarding AI grading for schoolwork, both before and after the task. After completing the task, students were randomly assigned to one of three conditions: being told they would be graded by an instructor, an AI, or an AI monitored by an instructor. Deception was used; this grading does not actually occur. Students' opinions and perceptions of distributive justice regarding AI grading for school were assessed.