Abstract:

Despite the ubiquitous use of generative artificial intelligence (AI), some educators are cautious about incorporating AI tools into their practices. Large Language Models (LLMs) are being adopted at all levels of education; however, usage of image generating neural networks (GANs and stable diffusion models) in education is understudied. Generative tools have the potential to drastically improve the quality of art education (Zhang et al., 2022). This paper aims to explore the construction of art teaching patterns involving image generation and to find out whether these patterns enhance or inhibit learning experiences in the context of art education.

We divided our research into two stages. In the first stage, we conducted a series of workshops where art educators were tasked with creating a lesson plan that incorporates image generating AI tools into their existing art educational workflows. Nine educational patterns were discovered by analyzing the resultant notes from the workshop using an inductive qualitative framework. In the second stage, the teaching patterns developed by art educators were adapted to learning scenarios for adult learners. A group interview was conducted with adult learners and learning facilitators to find out the value space of challenges and opportunities surrounding the teaching patterns. The value elicitation method can help to determine whether the use of AI tools in art education is meaningful and serves the well-being of the learning parties as well as to unravel the exposure to concerns and uncertainties (Väljataga et al., 2023).

Preliminary results from content analysis (Saldaña, 2012) show that the classical alexandrian pattern structure (problem - context - solution) might be enhanced by adding corresponding values and ethical dimensions. One such value: adaptability occurred during experimentation opportunities, where image generation tools allowed for adapting educational practices to the varying needs of students. Another value: involvement, which is proportional to the level of engagement, was perceived as crucial for ensuring meaningful learning experiences. Generative image tools were not unanimously received well; analysis showed some educators felt uncertainty about control over the AI tools in learning scenarios. In general, newly introduced AI inclusive teaching patterns were perceived as enhancing the autonomy, self-expression, and satisfaction of the learner.

Keywords:

Generative AI, educational patterns, value elicitation.