Paper Number

1913

Paper Type

Complete

Abstract

In order to achieve high-quality software, organizations use code reviews as an established quality assurance instrument in software development. With GenAI, organizations get the possibility to improve the time-consuming and cognitively demanding process of code review by assisting code reviewers with defect and improvement suggestions. As tools to support software development become more widely adopted, our research investigates the scenarios in which GenAI support helps reviewers perform better and addresses the relevance of mental workload in code reviews. As a result of our online experiment with 123 software developers, we can show that GenAI-assistance increases performance efficiency and effectiveness while decreasing the mental workload of reviewers when complex code is reviewed. This indicates that not only performance but also the human factor is important in the utilization of GenAI and the complexity of tasks should be considered.

Comments

11-PM

Share

COinS
 
Dec 15th, 12:00 AM

All Eyes on the Reviewer: Understanding the Impact of GenAI on Mental Workload and Performance in Code Reviews

In order to achieve high-quality software, organizations use code reviews as an established quality assurance instrument in software development. With GenAI, organizations get the possibility to improve the time-consuming and cognitively demanding process of code review by assisting code reviewers with defect and improvement suggestions. As tools to support software development become more widely adopted, our research investigates the scenarios in which GenAI support helps reviewers perform better and addresses the relevance of mental workload in code reviews. As a result of our online experiment with 123 software developers, we can show that GenAI-assistance increases performance efficiency and effectiveness while decreasing the mental workload of reviewers when complex code is reviewed. This indicates that not only performance but also the human factor is important in the utilization of GenAI and the complexity of tasks should be considered.

When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.