Understanding AI's Impact on Developer Workflows

Developers using AI coding assistants are unknowingly transforming their workflows in ways they can't even perceive. JetBrains researchers analyzed two years of behavior logs from 800 developers and discovered a striking disconnect between what programmers think is happening and what actually occurs in their daily coding patterns.

AI users typed significantly more code and deleted roughly 100 more lines per month compared to their non-AI counterparts, yet many reported no change in their editing habits. While developers felt their productivity and code quality improved, the behavioral data revealed they were actually debugging at similar rates and switching contexts just as frequently as before.

The study tracked everything from keystrokes to window switches, comparing 400 AI users against 400 non-users from October 2022 to October 2024. One developer with years of experience captured the complexity perfectly: "I triple-check it, and even then, I still feel a bit uneasy" about AI-generated code.

The research suggests AI tools are quietly reshaping how developers work rather than simply making them faster. These coding assistants redistribute effort across different activities, creating more iterative workflows where programmers write more code but also revise it more frequently, all while remaining largely unaware of these behavioral shifts in their own programming habits.