Teaching Novice Computing and Programming in the Agentic AI Era (cs.brown.edu)

A new viewpoint published in the Communications of the ACM proposes a transformative approach to computing education in light of advancements in AI programming agents. Authored by Kathi Fisler, Shriram Krishnamurthi, and Michael Littman, the paper suggests that instead of viewing AI as a disruptive force, educators should seize the opportunity to redefine introductory computing courses. By using the “flaky compiler” metaphor, the authors argue for a curriculum shift towards early software engineering principles, emphasizing the importance of responsibility and critical thinking in programming with AI tools. This significant shift in educational focus aims to better equip students to navigate the complexities of working with AI, ensuring they understand the implications of their code. The authors highlight the urgent need for research into innovative teaching methods for program specification and the integration of software engineering concepts in early computing education. As AI continues to evolve, rethinking how novice programmers are trained will be crucial not only for individual skill development but also for fostering a more responsible and effective AI landscape in the future.