AI Debugging Stumbles: New Study Highlights Challenges for Code Developers

A new study from Microsoft Research offers a sobering reality check to the hype surrounding AI’s coding capabilities. While advancements in AI are transforming software development, even sophisticated models still struggle with a crucial aspect of the process: code debugging. The study reveals that despite claims of AI generating up to 25% of new code at companies like Google and Meta, existing tools often fail to address complex issues that seasoned developers can easily resolve.