How to Critically Evaluate AI-Written Code?

AI coding agents can generate hundreds of lines of code in seconds — but should you trust what they write? In this video, JetBrains engineers put an AI coding agent to the test in real-world scenarios: analyzing dependencies in a Java project, implementing a bcrypt hashing function from scratch, and adding auditing support to a Spring application. Instead of asking whether AI can write code, we focus on the more important question: how do you critically evaluate it? You’ll see where AI is genuinely helpful and where it falls short — from exception handling and security concerns to architectural decisions and missing edge cases. Most importantly, we show why experienced developers are still essential, not to compete with AI, but to turn generated code into production-ready quality. AI is a powerful productivity tool — but it still needs a developer behind the wheel. What’s the most surprising AI-generated code you’ve reviewed? Let us know in the comments. Junie by JetBrains - https://www.jetbrains.com/junie/