A lot of this paper is dedicated to the usual description of how the 147 papers being surveyed were selected and classified, but there is some interesting discussion in what follows, including a section (4.2) on problems and AI-enabled learning interventions. Problems addressed include "difficulty sharing learning resources, the high redundancy of learning materials, learning isolation and inappropriate information load" as well as "high levels of demotivation, passive attitudes, boredom, poor engagement and frustration", but problems not addressed by AI interventions include "the use of outdated and highly complex models" as well as "personalisation issues, designing and assessing adaptive courses, high instructor workload" and more. And there's more to be done to actually make AI work. "Users do not understand how to extensively use such systems. At the same time, such systems—when implemented—have not actually overcome the complex challenges faced by students."
Today: 7 Total: 110 [Direct link] [Share]
