AI Says This Java Code Is Better. Let’s Check.

An AI assistant suggests a Java refactoring. The explanation sounds convincing. The code compiles. The tests pass. Would you merge it?

Your experience with code reviews, debugging, testing, and Java APIs gives you a strong starting point for evaluating AI-generated code. This session shows how to put that knowledge to work and choose tools that help you check the recommendations.

Drawing on examples from BYLES, Germany’s largest digital reading screening platform for primary school children, we investigate what makes a proposed change worth applying. You will learn how to approach verification, choose useful tools, and understand what the results tell you. Understanding this process also helps you identify the verification work you can later ask an AI agent to carry out.

The approach is independent of any particular AI assistant, IDE, or vendor. You will leave with practical methods for deciding what needs checking, choosing suitable tools, and using the evidence to make an informed merge decision. Build on what you already know and apply it with the tools you prefer.