Generative AI in Computer Science Education: Challenges and Opportunities

Description

Generative AI is a disruptive technology that has the potential to many aspects of how computer science is taught. Like previous innovations such as high-level programming languages and block-based programming languages, generative AI lowers the technical expertise necessary to create working programs, bringing the power of computation to more people. The programming process is already changing as a result of its presence, even for expert programmers. It also poses significant challenges to educators around rethinking assessment as some well-established approaches may no longer be viable. Many traditional programming assignments can be completed using generative AI tools with minimal effort, thus potentially undermining learning. In this Element, the authors explore both the opportunities and the challenges for computer science education resulting from the widespread availability of generative AI. Cover Title page Imprints page Generative AI in Computer Science Education Contents 1 Introduction 1.1 The Rise of Generative AI 1.2 Challenges and Opportunities in Education 1.3 Stucture and Audience 2 Understanding Large Language Models and ChatGPT 2.1 Image Classifier 2.2 Large Language Models 2.3 LLMs for Coding 3 Educator and Student Perceptions 3.1 Educator Perceptions 3.2 Student Perceptions 4 Class Preparation 4.1 Example Code Explanations 4.2 Worked Examples 4.3 Programming Exercises 4.4 Block-Based Programming Assignments 4.5 Assessment Questions 5 Class Instruction 5.1 Student Resources 5.2 Assignment Assistance 6 Designing for the User 6.1 Preparatory Materials 6.2 Integrated Development Environments (IDE) 6.3 Personalized Tutors 7 Challenges Posed by Learners' Misuse of Generative AI 7.1 Detecting Unauthorized Use 7.2 Learning Impacts from Misuse 7.3 Designing Courses to Avoid Misuse 8 New and Emerging Pedagogical Approaches 8.1 The Changing Industry Landscape 8.2 Writing and Explaining Code by Prompting 9 CS1-LLM: A Case Study of LLM Integration in an Introductory Coding Course 10 Conclusions References Acknowledgments