AI coding assistants (advanced)

The workshop AI Coding Assistants (Advanced), conducted by Leona Elisabeth Münzer and Sebastian David Schiller-Stoff at the University of Graz, explores AI-powered coding tools like GitHub Copilot, Codeium, and Google Colab’s integrated assistants. It provides hands-on demonstrations and exercises to showcase features such as autocomplete, chat-driven coding, and code refactoring.

Key focus areas include:

• Tool Comparison: Evaluating AI coding assistants for scripting, code suggestions, and debugging.
• Use Cases: Experiments demonstrate AI’s ability to assist in tasks like refactoring procedural code to object-oriented paradigms, efficient prompting, and managing “misleading context” problems.
• Efficiency: Participants engage in "TODO-driven software engineering," utilizing incremental prompts and interventions for improved workflow.

This session emphasizes the potential and limitations of AI coding assistants while fostering practical, critical judgment for developers at all skill levels.