Abstract:

Artificial Intelligence (AI) technologies in many areas, including education, are continuing to accelerate rapidly, in capabilities and availability. The level of AI hype and educators’ understandings of what AI is and does highlight the need for greater AI Literacy (Long & Magerko, 2020), along with AI Readiness, a contextualized way of helping people understand what AI is (Luckin, 2018), to support understandings of AI in Education (AIED), particularly ethical issues (Ng, Leung,, Chu & Qiao, 2021). With the introduction of ChatGPT in late 2022, increasing numbers of students and teachers are being exposed to AI technologies as everyday tools. When addressing the responsible use of AI, Dignum (2019) highlights the need for deep understandings of AI in context. Dignum (2021) claims that “the current educational system is not generating enough experts, and everyday users are increasingly not able to understand the systems with which they are faced” (p. 1). To better understand educators’ AI readiness and AI Literacy, we use mixed methods to examine the perspectives of graduate students’ interactions and the self-study reflections of the instructor’s in an Intro to AIED course taught in the Faculty of Education of a large urban Canadian university over four years, and more deeply examine these issues in a course on AI Ethics in Education.

The latter course is intended to:
a) develop learner understandings of what AI is,
b) move learners beyond simply implementing AI tools to thoughtfully considering integrating cultural, social, pedagogical, and ethical dimensions (Ouyang & Jiao, 2021; Pedro et al., 2019) and
c) apply perspectives of the usefulness of the TPACK (Technological Pedagogical Content Knowledge) framework and model card information to better understand the implications of contextually chosen AI tools in education.

Model cards are short documents accompanying trained machine learning models that provide benchmarked evaluation in a variety of conditions, specifically model details, typical use(s), known factors of use, metrics, training and evaluating data, ethical considerations, and caveats and recommendations (Mitchell et al., 2019). We applied AI classifications to help articulate these emergent learning processes, as such integrated frameworks within the education field are not currently developed. TPACK is a contextualized seven-item theoretical framework guiding the integration of technology into teaching (Mishra & Koehler, 2006). However, we found that while TPACK is a suitable framework for those in education who are incorporating traditional technologies, it is insufficient for identifying the challenges in learning and using AI tools, especially for those with limited technological backgrounds. On the AI side, model cards as a detailed tool for classifying AI technologies, supports the responsible democratization of AI and an increased transparency into their workings, specifically addressing ethical issues. We found the use of model cards requires learners to have developed significant levels of AI Literacy to be helpful to educators, yet their availability can support developing AI Readiness and AI Literacy. We will discuss our findings in greater detail in the full paper. We propose next steps for further research and explore an expansion of the TPACK framework to a similar framework within the AI context to support the ongoing development of educators’ AI Literacy and critical understandings of AI technologies.

Keywords:

AI Literacy, AI Ethics, AI Readiness, AIED, TPACK, Model Cards.