Development of evaluation tools for artificial intelligence educational AI platform

With the rapid introduction of AI technologies, there are increasing interest and expectations about the impact of AI in the education field. When a new technology is introduced, it is necessary to evaluate the potential and challenges of the technology from an objective stance. This study, hence, aims to derive implications by analyzing the current status and design level of educational AI applications in Korea and abroad. Specifically, the research questions examined in this study are: (1) what is the current status of the characteristics (subject area, target learner, AI technology type) of the educational AI applications?, (2) how can the educational AI applications be analyzed in terms of the level of complexity in AI design, and what is the current status of the design complexity levels? Through the systematic search process and selection criteria, 51 applications were selected as the targets to analyze their characteristics. In addition, focusing on the adaptive nature of AI technology, the level of design complexity was analyzed in terms of four levels according to the complexity and uncertainty of input-output. The main results for each research question are as follows. First, the number of applications developed for diverse groups was greater than that for single-group users. Second, there was some bias in the subject area and technology type of the analyzed applications. Concerning subject areas, many applications focus on language learning and mathematics learning, and speech recognition was the most frequently used AI technology. Third, the analysis of the complexity in AI design indicates that all applications were rather at low levels, namely ‘Level 1’ and ‘Level 2’. In conclusion, this study suggests that while social interest and investment in AI technology have increased recently, and the development of educational AI applications is increasing, AI applications that can be used in the actual classrooms are still limited and the level of intelligent functions may not be high, implying the danger of over expectations and hype toward AI in education.

DESIGN COMPLEXITY OF MOBILE LANGUAGE APPS

LEARNING MECHANISM

References

2020

1. KCI

   Analysis of the characteristics and design level of educational artificial intelligence applications

   Hyeran Lee, Hyo-Jeong So, and Lingxi Jin

   Journal of Korean Association for Educational Information and Media, 2020

   DOI