Vol. 4, Issue 6 (2019)
Artificial intelligence and voice-powered electronic textbooks
Author(s): John Leddo, Yihao Guo, Yuwei Liang, Rushil Joshi, Ivy Liang, Wei Guo, Stacey Bailey
Abstract: Textbooks have long been an integral part of classroom education. As much of the information world has moved online, so, too, have textbooks followed. Unfortunately, commercially-available electronic textbooks (e-textbooks) are typically little more than .pdf versions of their paper counterparts, thus not exploiting other technologies that could be used to increase student learning. The present paper describes e-textbook technology that uses artificial intelligence (AI) and voice/natural language technologies to increase student learning. As students are learning a lesson, they can verbally ask the e-textbook questions about the topic and receive answers much the way they can when using personal assistants on smart phones. When students have completed the lesson, the e-textbook assesses whether they have learned the material by verbally asking questions and allowing students to answer verbally. Any deficiencies are immediately remediated. When students finish the assessment, they do practice problems as they would in a standard e-textbook. The difference is that with the present technology, all work is done step-by-step on an electronic worksheet where the underlying AI technology monitors each step and provides hints when requested and feedback when mistakes are made. The present technology was tested experimentally by having students either use it or leading publisher Pearson’s Algebra 2 Common Core e-textbook to learn division of complex numbers. Students were then given a post-test to measure their learning. Students using the AI and voice/natural language-powered e-textbook scored four times high on the post-test than those using Pearson’s e-textbook. The results suggest that AI and voice/natural language technologies can improve educational performance when incorporated into e-textbooks.