The study also found that churned code, or work that was corrected within two weeks, increased from 3.3% in 2020 to 5.5% in 2023, with a forecast of 7.1% for 2024. The proportion of developers using AI assistants also rose to 30% in 2023. The study suggests that the use of AI assistants is leading to more incorrect code being pushed into repositories and corrected later, negating the 55% efficiency gain promoted by GitHub. The study recommends the development of special cleaning tools to minimize this technical debt.
Key takeaways:
- A study by GitClear suggests that the increasing use of AI assistants like GitHub Copilot and ChatGPT is leading to a decrease in code quality and an increase in technical debt.
- The study analyzed 153 million lines of code and found that the use of these AI assistants often results in quick but not necessarily good and sustainable solutions, leading to code repetition and increased maintenance effort.
- The proportion of churned code, or work that was corrected within two weeks, increased from 3.3 percent in 2020 to 5.5 percent in 2023, and is forecasted to reach 7.1 percent in 2024.
- The study recommends the development of special cleaning tools to minimize this technical debt and questions whether there are any AI assistants trying to solve the problem.