Abstract

In this presentation, I explore the opportunities that generative AI (GenAI) opens for computer science (CS) and data science (DS) education. I argue that GenAI not only does not eliminate the need for CS and DS education, by eliminating the need to deal with the technical details of programming and data analysis at some stages of the learning process, GenAI actually enables us to increase the level of abstraction and complexity of the tasks that we assign students and of the skills and competencies that we seek to impart.
Specifically, to illustrate the horizons that GenAI opens for CS and DS education, I revisit pedagogical and cognitive models and theories, and explore their implementation in the GenAI era in the context of CS and DS education. These theories are categorized into three groups: learning, pedagogy, and content/competences:
1. Learning
1. Constructivism (Ben-Ari, 1998; Papert, 1980)
2. Cognitive load (Sweller & Chandler, 1991)
3. Motivation (Deci & Ryan, 2000; Ryan & Deci, 2017)
2. Pedagogy
1. Blum’s taxonomy (Bloom et al., 1956)
2. Assessment
3. Personalization/diversity/equity
4. Didactic transposition (Chevallard, 1985)
3. Content / Competencies
1. Knowledge, skills, attitudes (KSA) model
2. Computational thinking (Wing, 2006)
3. Metacognition
For example:
• With respect to learning, I refer to cognitive load (Sweller & Chandler, 1991) and illustrate how learning processes that are accompanied by GenAI tools can manage intrinsic cognitive load more effectively, reduce extraneous cognitive load, and promote germane cognitive load – the cognitive load that leads to meaningful learning processes.
• When talking about pedagogy, I focus on didactic transposition (Chevallard, 1985), which refers to the transformation of expert knowledge into teachable knowledge, that is, adapting complex or specialized content so that it can be transmitted effectively to students. I demonstrate how GenAI can help educators deductively transpose a complex machine learning concept (such as back propagation) into teachable material for any level or age group.
• Finally, with respect to content/competencies, such as computational thinking (Wing, 2006) and metacognitive skills, I argue that not only do these skills increase in importance, but as the use of GenAI becomes more and more widespread, computational thinking and metacognitive skills become essential skill for all. Specifically, in the GenAI era, computational thinking manifests in a new form, namely prompt engineering, which all GenAI users are required to master (Erez, Mike & Hazzan, submitted).
One important takeaway from this presentation is that the CS and DS education community should perceive GenAI not as a threat, but rather as an opportunity to achieve pedagogical targets and transmit the values that we have always sought to impart, but could not due to the need to deal with technical details involved in programming and data analysis.