GenAI in Software Engineering: The Role of Technology Acceptance Models
Oscar Johansson, Jürgen Börstler, Nauman bin Ali
TLDR
This paper proposes enhancing UTAUT with Bayesian analysis to identify and overcome individual barriers to GenAI adoption in software engineering.
Key contributions
- Refine UTAUT constructs to better capture GenAI's unique nature and transformational impact.
- Improve operationalization practices for stronger construct validity and cross-study comparability.
- Integrate Bayesian analysis for small-sample inference, prior knowledge, and iterative model updating.
Why it matters
Many organizations are eager to integrate GenAI into software development, but individual acceptance barriers can hinder adoption. This paper provides a framework, using UTAUT and Bayesian analysis, to systematically identify and address these challenges, paving the way for more effective GenAI implementation in SE.
Original Abstract
Context: Many organizations are keen to incorporate generative~AI (GenAI) into their software development processes. Technology acceptance models, such as the Unified Theory of Acceptance and Use of Technology (UTAUT), are traditionally used to identify individual-level barriers to the acceptance of new technologies and can facilitate the transition to GenAI. However, UTAUT has seen limited use within software engineering (SE) research. Objective: Using UTAUT as an example, to identify key areas for future research on GenAI acceptance, including the role of Bayesian approaches for data analysis. Method: We review foundational and SE-specific literature on UTAUT and analyze its emerging applications for GenAI in SE. Results: We identify three priorities for future research: (1) identifying and refining constructs to account for GenAI's nature and transformational impact; (2) improving operationalization practices to strengthen construct validity and cross-study comparability; and (3) incorporating Bayesian analysis to support small-sample inference by integrating prior knowledge, iterative model updating, and simulation of scenarios. Conclusion: UTAUT is a suitable candidate to combine with Bayesian analysis for practical insights on individual-level barriers to GenAI use in SE, but additional theories should be considered.
📬 Weekly AI Paper Digest
Get the top 10 AI/ML arXiv papers from the week — summarized, scored, and delivered to your inbox every Monday.