Navigating the Complexity of Generative AI Adoption in Software Engineering
Осмысление сложности внедрения генеративного искусственного интеллекта в разработку программного обеспечения
2024-03-28
SCID: 54.1/n2thr54e
Discuss with AI
Generative AI adoptionHuman-AI Collaboration and Adaptation FrameworkPartial Least Squares–Structural Equation ModelingSoftware engineeringTechnology Acceptance Model
Figures from the paper
Abstract (AI)
This article explores the adoption of Generative Artificial Intelligence (AI) tools within the domain of software engineering, focusing on the influencing factors at the individual, technological, and social levels. We applied a convergent mixed-methods approach to offer a comprehensive understanding of AI adoption dynamics. We initially conducted a questionnaire survey with 100 software engineers, drawing upon the Technology Acceptance Model, the Diffusion of Innovation Theory, and the Social Cognitive Theory as guiding theoretical frameworks. Employing the Gioia methodology, we derived a theoretical model of AI adoption in software engineering: the Human-AI Collaboration and Adaptation Framework. This model was then validated using Partial Least Squares–Structural Equation Modeling based on data from 183 software engineers. Findings indicate that at this early stage of AI integration, the compatibility of AI tools within existing development workflows predominantly drives their adoption, challenging conventional technology acceptance theories. The impact of perceived usefulness, social factors, and personal innovativeness seems less pronounced than expected. The study provides crucial insights for future AI tool design and offers a framework for developing effective organizational implementation strategies.
Key Findings
1
Compatibility with existing software-development workflows predominantly drives early Generative AI adoption, challenging conventional technology acceptance assumptions.
2
Perceived usefulness, social influences, and personal innovativeness have less influence on adoption than expected, informing tool design and organizational implementation strategies.
3
The framework integrates individual, technological, and social influences using a convergent mixed-methods design and three established adoption theories.
4
The study develops the Human-AI Collaboration and Adaptation Framework to explain Generative AI adoption in software engineering.
5
Validation with Partial Least Squares–Structural Equation Modeling used data from 183 software engineers after an initial survey of 100 engineers.
Research Object
Generative AI tools in software engineering
Research Subject
Factors and dynamics influencing the adoption of Generative AI tools, particularly their compatibility with existing software development workflows
Publication Details
Publication Date
2024-03-28
Journal
Publisher
ISSN
Cited by
217
Open access PDF
Access Type
Author Information
Download PDF
Subscribe to digest
References available in scid.ai8
Perceived Usefulness, Perceived Ease of Use, and User Acceptance of Information Technology1989
User Acceptance of Computer Technology: A Comparison of Two Theoretical Models1989
The Discovery of Grounded Theory: Strategies for Qualitative Research2017
A Theoretical Extension of the Technology Acceptance Model: Four Longitudinal Field Studies2000
Computer Self-Efficacy: Development of a Measure and Initial Test1995
Scaling Laws for Neural Language Models2020
Predictors of well-being and productivity among software professionals during the COVID-19 pandemic – a longitudinal study2021
The Daily Life of Software Engineers During the COVID-19 Pandemic2021