Year: 2021
Type: article
Abstract: Current educational assessments report that, nationally, the majority of United States students are performing below grade level (National Assessment of Educational Progress, 2019).Because this trend has been ongoing and there is a current teacher shortage, it seems unlikely that this problem will be rectified in the foreseeable future unless a paradigm shift occurs.The present paper explores the question of whether using artificial intelligence (AI)-powered educational software can effectively stand in for human teachers.High school students were taught the difficult Algebra II topic of dividing complex numbers.Half were taught by active high school math teachers while half were taught by AI-powered software that our team created.Results showed that, on a post-test, students using the AI-powered software scored, on average, 37% higher than those taught by teachers.A closer examination of the data reveals that the distribution of scores of students taught by teachers is rectangular with some performing highly, some performing in the average range and some failing, while most students who were taught by the AI software scored in the A range and no one scored below 70%.Results suggest that our AI software could potentially supplement or stand in for teacher-led instruction with an overall improvement in student performance.
Authors: John Leddo, Khushi Garg
Language: English
Subfield: Education
Field: Social Sciences
Domain: Social Sciences
SDG: Quality education
Open Access status: diamond
OpenAlex ID: W4206748396