Copilot, not autopilot: A hands-on workshop for responsible AI in dev workflows

AI tools can accelerate web development work, but only if developers use them with judgment. This hands-on workshop is designed for web professionals who want to use tools like GitHub Copilot, ChatGPT and Claude more effectively without outsourcing their thinking. Rather than treating AI as an autopilot, you will practice using it as a copilot: useful for drafting, debugging, reviewing and exploring solutions, but always subject to human evaluation.

Through guided exercises, you will work through realistic development scenarios that show where AI can help, where it can mislead and how to review its output responsibly. You will practice choosing appropriate use cases for AI assistance, improving prompts with better context, evaluating generated code for correctness and maintainability, using AI to support debugging, and comparing AI-generated pull request feedback against your own engineering judgment.

The workshop is built as a supervised lab, not a lecture. You will actively use AI tools during the session, discuss your results in small groups and receive guidance from the workshop leader as you work. By the end, you will leave with practical methods you can apply in your own development workflows, along with a clearer sense of when AI speeds things up, when it introduces risk and how to keep human understanding at the center of the work.

By the end of this workshop, you will be able to:

• Identify which development tasks are good candidates for AI assistance and which require greater caution or should remain primarily human-driven.
• Write more effective prompts by supplying the technical, contextual and quality constraints AI tools need to produce useful output.
• Evaluate AI-generated code for correctness, maintainability, accessibility, security and fit within a real codebase.
• Use AI as a support tool for debugging and pull request review without replacing human engineering judgment.
• Define a practical set of personal or team guardrails for responsible AI use in day-to-day development workflows.