미술감상교육에서 생성형 AI의 활용가능성

Abstract

Recently, a variety of generative AIs have been commercialized. There are many concerns about AI, but there are also efforts to use it in education. Generative AIs such as Midjourney are being used to produce images, and there is also public discussion about this. In addition, in art appreciation education, the conventional method of writing appreciation essays can be utilized through language-generating AIs such as ChatGPT. In this article, we applied this to the writing of an appreciation essay on the Korean traditional Chaesekhwa <Ilwolobongdo>, which has relatively little information. ChatGPT wrote an appreciation essay based on incorrect information about the work being analyzed. The less information it provided in the question, the falser writing it wrote. On the other hand, Bard and Bing used the search engines that were associated with each other to answer with more accurate information. As a result, the more specific the direction of the question, the more relevant the writing was. There were clear limitations, such as providing false information or writing an appreciation with insufficient evidence. However, I think that in the future, we will be able to write a text based on accurate information if it is done in parallel with the search and more information is accumulated. Accordingly, in art appreciation education, it will be possible to do critical appreciation education based on the contents that the generative AI has found and organized, rather than focusing on writing appreciation essays.